Gamma distribution model (GDM) and artificial neural network (ANN) have been used to predict the weekly rainfall probabilities of Anand station of Gujarat , India using 48 years of rainfall data series (1958 to 2005). Estimated probabilities by GDM were compared with actual probabilities. Artificial neural network was used with back propagation algorithm and it was trained with the probabilities (%) obtained by GDM for weekly rainfall of 0.25cm and 0.5cm. Parameters used to train the neural network were number of hidden neurons 140, momentum 0.5 and error goal (computer error) 10 -22 . Probabilities obtained by ANN for different amounts of weekly rainfall were compared with probabilities obtained by GDM. The probabilities computed by both the methods GDM and ANN for getting certain fixed amount of rainfall were significant to actual probabilites. All the related programmes were developed in MATLAB.