This study is an endeavor to determine the relationship of some of the selected climatic variables and a non-climatic variable with Aman rice yield estimation. The study is also intended to forecast the yield of the Aman crop and to determine the difference between the forecast and actual yield estimation by varieties at different stages of the crop of Rangpur district in Bangladesh. On the basis of forecast estimate pertinent authority could take appropriate actions such that the sufferings of the rural people are minimized. Data were collected from district meteorological offices and district fertilizer sales office of Rangpur district. Data were collected from 1968 to 1988 in the weekly form Forecasting equations were formulated where predictor variables showed significant effect at different stages (i.e. vegetative reproductive, grain filling and maturation stages) for Aman rice crop on weekly basis. Simple and multiple linear regression analysis were used for analytical purposes. F-test, T-test and test of lack of fit were used to test the validity of the models. Statistical analysis which showed retrodictive results (20 years average) very close to actual estimate of Rangpur district were IIYV Aman at Vegetative and at reproductive stages whereas for LV Aman at Vegetative, reproductive, grain filling and maturation stage. Recommendations were made for inclusion of more non weather variables and to use the models for other districts for forecasting.