The manual diagnosis for the diseases of agricultural crops is often restricted by the individual ability and
experiences so that one cannot obtain the precise results of diagnosis. To overcome this pitfall, the merge of expert systems
with the rich pathological knowledge and the utilization of pattern recognition algorithm can significantly improve the
precision of diagnosis. Therefore, it greatly increases the quantity and the quality of crop’s production. In this paper, a
pattern recognition technique, based on a radial basis function (RBF) neural network is applied to the diagnosis of soybean
diseases. The RBF neural network, which is a novel and efficient feed-forward network, is based on the local reflections of
cortical neurons on the external stimulus. This network possesses variety of characteristics, for example, the simple
structure, strong global convergence, and fast-speed training behavior, which together make RBF network to be used widely in
the field of pattern recognition. Firstly, nineteen typical symptoms of soybean diseases are collected, and then constructed
to form an experimental sample set. Secondly,a RBF network model is set up and trained using algorithm programming. Finally,
the result of test shows that the RBF model is of high diagnosis precision and strong generalization ability.