. In Indonesia, agriculture becomes an important sector for national development and national economy. The onset of the rainy season is one of the rainfall variables that affect agricultural production. The changing of the onset of rainy season can impact on crop failure. This research aims to develop a model for predicting the onset of rainy season using optimized cascade neural network with genetic algorithm based on global circulation model in Pacitan district. Observational data used is the onset of rainy season of 3 weather stations in Pacitan: Arjosari, Kebon Agung, and Pringkuku. Predictor data used is global circulation model output data between 1983 – 2011 from 3 models: CMC1-CanCM3, CMC1-CanCM4, and NCEP-CSFv2. Optimization of cascade neural network with genetic algorithm has been done by optimizing the amount of hidden neuron and obtained an increase value of correlation coefficient (r). This research obtained the best model from each weather stations with different parameters. R value of Arjosari weather station is 0.89. R value of Kebon Agung weather station is 0.86. R value of Pringkuku weather station is 0.87.