An artificial neural network is applied to the operation control of the photovoltaic/diesel hybrid power generation system. The optimal operation patterns of the diesel generator are calculated by dynamic programming (DP) under the known insolation and load demand, which minimize the fuel consumption of the diesel generator. These optimal patterns are learned by the three layer neural network, and it is tested for the different insolation and demand data from those used in the learning. Two kinds of neural networks are examined, and the results are compared with each other.<<ETX>>