In this paper we present our Bayesian method for carrying out motor unit number estimation (MUNE) using stimulus–response data collected from surface electrophysiological recordings. We formulate and justify our assumptions in Ridall et al. (2006) and base these on available scientific evidence, as outlined in the previous paper. The object of our methodology is to express the uncertainty about the number of motor units in a muscle in terms of a posterior distribution. From studies taken over time, these posterior distributions can be used to estimate the rate of loss of motor units. A by-product of this method of MUNE is that it provides a means of tracking a given population of motor units over time. Examples of the parameters that can be tracked over time include the distribution of the excitability parameters, the single motor unit action potentials and the between and within motor unit variability.