A turbulent combustion model based on the probability density function (PDF) approach has been extended for the spray combustion computations under simulated diesel engine conditions. This approach accounts for the effects of turbulence and of random dynamics of evaporating droplets on the mean rate of chemistry. The paper begins with formulation of the equation for the joint PDF for the gaseous mixture variables where evaporating droplets are viewed as point sources. Then the modified micromixing model that involves evaporation process is described and details of a Monte-Carlo modeling of PDF-equation are given. After that, the results of numerical studies dealing with diesel spray combustion are discussed. Three different examples are considered. First, the computations of the evaporating spray injected into heated nitrogen atmosphere are carried out. The contribution of evaporating droplets dynamics to the mean and variance distributions of temperature and vapor concentration is demonstrated. Next, the spray combustion under light-duty and heavy-duty diesel conditions is simulated, and the results are compared with calculations using the Eddy-Break-Up combustion model, and with experimental data. It is shown that the PDF equation model is able to predict experimental data significantly better than the Eddy-Break-Up model. The last part of computations concerns the diesel spray autoignition governed by the strong turbulence effects. Spatial probability distributions of self-ignition sites are displayed and compared with experimental observation. It is shown that the region with high probability of ignition sites occurs at the level of dense spray and is displaced towards the nozzle hole if the inlet air temperature and pressure increase.