You do not actually need to use photons here - in fact the often sadly-ignored in many people's minds good ole' classical description of light as an EM wave is sufficient (though that doesn't mean the other way is wrong, but it's important to point it out!).
A light wave, like any wave, has three basic parameters: its amplitude, frequency, and wavelength. (You may also add phase, but phase is really just a reference shift, it does not change the shape of the wave, though it is important when summing waves.) The amplitude of a light wave is the maximum size of its electric field vector, just as the amplitude of a sound wave is the maximum strength of the overpressure relative to the ambient pressure, and the amplitude of a sea wave is the maximum height of a wave crest above the surrounding unperturbed sea level. The amplitude determines the amount of energy, and thus the intensity - what we'd call respectively "brightness", "loudness", and "height", that is transmitted by each wave.
Amplitude has nothing to do with speed - rather speed is what relates frequency and wavelength: $\lambda f = v_w$, where $v_w$ is the speed of the wave. Amplitude ($A$) does not appear in this relation. That means the intensity can go up, down, or stay the same, and have no effect on the speed. Here, of course, $v_w = c$, the speed of light waves in vacuum, or $v_w = \frac{c}{n}$ in a refractive medium with refractive index $n$.
In your scenario, as the waves spread out, their amplitude decreases - just as a sound wave gets less loud the further you go, and a water wave shallows as it travels out from a source of disturbance such as a cast rock. Because speed has nothing to do with amplitude, this can happen without requiring any changes in speed.
In terms of photons, if you want to use them, the intensity is related to the photon number together with frequency. Decreasing the intensity reduces the number of photons, but each photon always travels with speed $c$. In this scenario, the photons become more thinly spread, over a wider region, so a smaller number in each volume of space.
(And yes I'm assuming a "linear" situation in the above, yes caveats about nonlinearity, yes yes, blah blah, but for these purposes, that is sufficient.)