Based on discussing the advantages of improving genetic algorithm and analyzing the defects of back propagation neural network, we presented the genetic neural model. The simulating data proved that the genetic neural model was able to realize parallel search and could get faster searching speed during random searching optimizaiton. The model was applied to predicting distribution of guava fruit fly. The experimental results show that the model can predict distribution of the fly which is consistent with the practical distribution. The model is very useful in practice. It is worthwhile to refer the model to predicting similar insects.