It's in Lorentz' 1895 paper:
"Versuch einer theorie der electrischen und optischen erscheinungen
bewegten kõrpern"
which you can find copies of on the net. It's buried under notation alien to current eyes, and is in German. I can read German, but it's a drag for me to do so.
Here's an English translation:
Attempt of a Theory of Electrical and Optical Phenomena in Moving Bodies
https://en.wikisource.org/wiki/Translation:Attempt_of_a_Theory_of_Electrical_and_Optical_Phenomena_in_Moving_Bodies
I can't vouch for its accuracy.
The "moving" - as in Einstein's 1905 paper - refers specifically to the condition $ ≠ $ (or, in his notation: $ ≠ 0$) of being in a "non-stationary" frame. That's discussed in further detail below.
He mentions the force in §12 of Abschnitt I; lays out some equations - for empty space - in that section, and the equations for moving media in Abschnitt II. You need a Rosetta Stone to penetrate the text. So, for that, let's lay out a few points of reference.
He apparently expressed the Lorentz force, there and later, as a force $_1 = + ×$ per unit charge, rather than as a force outright $ = e( + ×)$. Part of this was already there with the velocity-correction $ + ×$ that is already present in the constitutive law for the electric field - see below. The velocity $$ is relative to the "stationary" frame. Again: see below. So adding on the extra velocity term $( - )×$ for the velocity $$ of the moving charge is a natural step - and a step required to make the force law independent of any specific frame.
Our current understanding includes the Maxwell equations:
$$∇·, \quad ∇× + \frac{∂}{∂t} = , \quad ∇· = ρ, \quad ∇× - \frac{∂}{∂t} = .$$
This holds across the board, relativistically and non-relativistically as is, because they live at a deeper layer of geometry in which such distinctions (even distinctions between space-like and time-like) are not present.
It was also common then - and now - to define the "total current":
$$ = + \frac{∂}{∂t},$$
so that one could just as well write $∇× = $. The current $$ is the one used in circuit analysis, not $$.
Our understanding, at present, also includes the following constitutive relations:
$$ + α× = ε( + ×), \quad - α× = μ( - ×).$$
For Relativity, $α = (1/c)^2$ and these are the Maxwell-Minkowskii relations. For non-relativistic theory $α = 0$, and they are the relations posed by Maxwell with the correction $-×$ later added by Thomson.
The velocity $$ is relative to the frame in which the constitutive laws assume isotropic form $ = ε$ and $ = μ$. In that frame, $ = $, and may be referred to as the "stationary" frame.
The coefficients $ε$ and $μ$ are properties of the medium, with $εμ > 0$; the speed $V = 1/\sqrt{εμ}$ is a characteristic of the medium and determines the speed of wave propagation, including that of light - relative to the stationary frame.
In the case where $εμ = α$, provided that $|| < V$, all frames are equivalent to the stationary frame, the Maxwell-Minkowski relations are equivalent to their $ = $ version, for all $$, and $$ drops out of the picture. In that case, $V = c$, the invariant speed mandated by Special Relativity. These are the relations for the vacuum. Hence, $c$ is referred to as the speed of light in a vacuum.
The equations laid out by Lorentz in Abschnitt II are:
$$\begin{align}
Ⅰ_b\ & \text{Div}\ = ρ \\
Ⅱ_b\ & \text{Div}\ ℌ = 0 \\
Ⅲ_b\ & \text{Rot}\ ℌ' = 4πρ + 4π\dot{} \\
Ⅳ_b\ & \text{Rot}\ = -\dot{ℌ} \\
Ⅴ_b\ & = 4πV^2 + [·ℌ] \\
Ⅵ_b\ & ℌ' = ℌ - 4π[·] \\
Ⅶ_b\ & = + [·ℌ]
\end{align}$$
In other contexts, in Abschnitt I, he also makes mention of the convection current $ℭ = ρ$ and the total current $ = ℭ + \dot{}$. He also uses the notation
$$Δ = \left(\frac{∂}{∂x}\right)^2 + \left(\frac{∂}{∂y}\right)^2 + \left(\frac{∂}{∂z}\right)^2$$
and $Δ'$ for $Δ$, with the effect of the velocity $$ subtracted off, and $(∂/∂t)_1$ being $∂/∂t$ with the effect of the velocity $$ subtracted off. His $\dot{(\_)}$, going by context, is equal to our (and his) $∂/∂t$, while his "Div" is equivalent to our $∇·(\_)$ and his "Rot" to our $∇×(\_)$, where
$$∇ = \left(\frac{∂}{∂x}, \frac{∂}{∂y}, \frac{∂}{∂z}\right).$$
His vector product $[\_·\_]$ is what we today write as the vector cross product $(\_)×(\_)$.
The Rosetta Stone, with Lorentz on top, and our notation on the bottom, is:
$$
\left(ℌ, , , ℌ', , ρ, ℭ, , , \right) \\
⇒ \\
\left(\frac{}m, \frac{}m, \frac{_1}m, 4πm, m, mρ, m, m, , -\right)
$$
for the objects, where $m = \sqrt{μ/(4π)}$; and
$$
\left(Δ, Δ', \text{Div}, \text{Rot}, \dot{\left(\_\right)}, \left(\frac{∂}{∂t}\right)_1, [\_·\_]\right) \\
⇒ \\
\left(∇², ∇² - \frac{(·∇)^2}{V^2}, ∇·(\_), ∇×(\_), \frac{∂}{∂t}, \frac{∂}{∂t} + ·∇, (\_)×(\_)\right)
$$
for the operators. The specific factors used are those required to also reconcile the energy integrals, which were cited elsewhere in the article.
With these correspondences, his equations become:
$$\begin{align}
Ⅰ_b\ & ∇· = ρ \\
Ⅱ_b\ & ∇· = 0 \\
Ⅲ_b\ & ∇× = + \frac{∂}{∂t} \\
Ⅳ_b\ & ∇× = -\frac{∂}{∂t} \\
Ⅴ_b\ & = /ε - × \\
Ⅵ_b\ & = /μ + × \\
Ⅶ_b\ & ₁ = + ×
\end{align}$$
with the constitutive law $ = ρ$ for the current.
So, if he originated this, then his contribution was to take the force law for the "moving" electric force $ + ×$ and to then make it frame-independent, by generalizing the $$ in the $×$ term to $$. Maxwell's $$, in fact, was actually already defined as the "moving" electric force, not as our $$. It muddied up his equations and formalism.
As you can see ... by the way ... Lorentz' equations are a combination of Maxwell's equations plus the non-relativistic version of the constitutive law, along with the per-unit-charge force law and the decomposition law $ = ρ$ for the current. Lorentz' theory was firmly couched in non-relativistic physics because of the absence of the relativistic corrections $+ α×$ and $- α×$ with $α = (1/c)^2$ present in the Maxwell-Minkowski relations.
