In this paper, an analysis of volumetric efficiency of hydrostatic pumps in a variety conditions is investigated by using a proposed neural network. The effects of the parameters, such as the number of revolution, hydraulic oil temperature and exit pressures, which act on performances of hydrostatic pumps like gear pumps, vane pumps, and axial reciprocal pumps with swash plate, on the volumetric efficiency have been examined. The revolution number of the pumps, exit pressure of the system and the hydraulic oil temperatures are greatly affected by the leakage flowrate. The neural network structure is very suitable for this kind of system. The network is capable of predicting the leakage flowrate of the experimental system. The network has a parallel structure and fast learning capacity. As it can be seen from the results for both approaches, neural network could be modeled hydrostatic pump systems in real time applications.