In this paper, a three-dimensional stochastic model is extended to simulate the thermal spray coating process from droplet impact to coating formation on a solid surface. Process parameters of molten droplets landing on the substrate are generated randomly by assuming that these properties follow normal distributions with user-specified means and standard deviations; an empirical correlation is applied to relate particle velocity V to diameter D and Temperature T. Splat sizes after droplet impact are calculated from an analytical model and previous rules of droplets interactions are improved. Porosity is assumed to be caused by the curl-up of splats as a result of thermal stresses. We use a Cartesian grid to define the computational domain and to track the shape and position of the deposited coating. The surface of the coating and the location of pores within it are specified using a variable known as the “volume fraction”. The model is capable of predicting the variation of coating porosity, average thickness and roughness as a function of process parameters. Simulated predictions agree with experimental observation.