During last decades the environmental and economic effects of climate changes have been very important for environmental scientists. According to temperature fluctuation studies, the minimum, maximum, and average temperatures are to be increased. So, the objective of this study is the investigation of variations of the number of freezing days during a 57 years data set and modeling the data by multivariable statistics methods. The applied models were ARIMA time-series, Regressions, and Correlation coefficient. The results showed that the variations of temperature were very disorder from 1951 to 1975 and they were not following a specific model. However, it showed the decreasing in freezing days during 1976-2007 which could be due to urban warming and crating island effect in this duration.