Precipitation is one of the key elements in the climate of each region. Decrease or increase precipitation, climate and environmental phenomena that impress direct effects of climate on human life are significant. Statistical techniques are a useful tool for predicting the behavior of the climate variables. In this study, using statistical methods, precipitation behavior is analyzed in Khoi meteorological station. For this purpose, statistical data of annual average Precipitation during the period 960- 2011 have been used. Using the methods like Pearson, Spearman and Man-Kendal, we have attempted to investigate the precipitation trend. The results of application of these methods show significant decreasing trend in annual rainfall in Khoi meteorological station. By applying a spectral analysis method based on precipitation data, its full cycle, was evaluated. The results of spectral analysis showed that at 95% confidence level, the first harmony was significant. Finally the Arima model was used to predict annual precipitation in the study area. Four basic models were fitted. Goodness of fit tests, including tests of coefficients, remained independent test of the model, using Akaike and prediction model, indicates that between the four models fitted, Arima model (1, 1, 0) is the best fitted to annual precipitation. Based on this model, Khoi meteorological station annual precipitation was predicted in 95 percent level by 2016.