Cell migration is a vital process in living organisms. In particular we are interested in the way that white blood cells such as neutrophils migrate during episodes of inflammation which are important events in the working of the innate immune system. Migration of populations of many kinds can be modelled using drift-diffusion models by drawing analogies between the individual agents and the molecules in a fluid. It is challenging to arrive at a data-driven estimate of the parameters of this kind of process, particularly so if the individual agents have time varying properties that are not uniform over the population. In this paper, we offer a novel framework to estimate migration dynamics in this context. It makes use of the Approximate Bayesian Computation approach for parameter estimation and model selection. The Framework is applied to zebrafish neutrophil dynamics but is applicable for general migration scenarios.