The forecasting of the rainfall during December until May in Khorasan province (Northeast of Iran) by use of Adaptive Neuro – Fuzzy Inference System and Mamdani Fuzzy Inference System has been illustrated .This method includes three steps. The relation between the changes of Synoptic patterns including sea surface pressure , the sea surface pressure gradient, sea surface temperature , difference of sea surface temperature and 1000 mb, the temperature at 850 hpa, height of 500 hpa level and relative humidity of 300 hpa with local average rainfall had been considered in the first step . The rainfall impression from synoptic patterns was noticeable in the North- east of Iran in choosing this areas which include a collection of localities in Persian gulf and Oman sea ,Black sea, Caspian Sea , Mediterranean sea , North sea , Adriatic sea , red sea, Adan gulf ,pacific Ocean, Indian and Siberian Oceans. In the second step the model had been taught during 1970-1993 and in the third step, the rainfall forecasting had been done during 1993-2002. The results show that the Adaptive Neuro-Fuzzy Inference System can predict the rainfall with acceptable precision in 70% of the years.

Keywords