Drought is one of the most important natural disasters that is effective in different parts such as; natural resources, surface and subsurface water resourcesand etc. Notice to the changes trend of drought and its forecasting, can be effective in planning of drought control and reduction of drought impact. In this study climate data of 7 stations of southwest of Iran from 1980 to 2014 was used and the Intensity of seasonal drought was calculated based on RDI index. Then drought index predicted from 2015 to 2019 using time series models and ITSM software, then the changes trend of drought based on parametricand non-parametric tests were evaluated. According to results in Abadeh, Shiarz and Fasa stations MA (5) Innovation, in Bandar Abass station AR (15) Burg, in Bandar Lengeh station AR (15) Yule-Walker, in Boushehr station MA (5) Hannan-Rissanen and in Jask station AR (13) Burg model were the best time series model that fitted in data. Correlation coefficient between observed data and predicted data, P- Value of Ljung – Box test, white noise of models and ACFand PACFResidual of models showed that predicted models were Reliable. Results showed that the changes trend of seasonal droughtin all stations from 1980 to 2014 and 1980 to 2019 periods were non-significant at 0.95 level based on parametric and non parametric statistical methods. However slop of changes trend in (1980-2019) period was more than (1980-2014) period.