Despite of the small amount in the atmosphere, ozone is one of the most critical atmospheric component as it protects human beings and any other life on the earth from the sun's high frequency ultraviolet radiation. In recent decades, the global ozone depletion caused by human activities is well known and produces an "ozone hole", the most direct consequence of which is the increase in ultraviolet radiation, which will affect human survival, climatic environment, ecological environment and other important adverse impacts. Due to the implementation of the Montreal protocol and other agreement, the total amount of ozone depleting substance in the atmosphere has been prominent reduced, which will lead to a new round of regional climate change. Therefore, predicting the changes of the total ozone in the future will have an important guiding significance for predicting the future climate change and making reasonable measures to deal with the climate change. In this paper, based on the ozone data of 1979 to 2016 in the southern hemisphere and ARIMA model algorithm, using time series analysis, we obtain prediction effect of ARIMA model is good by Ljung-Box Q-test and , and the model can be used to predict the future ozone change. With the help of SPSS software, the future trend of the total ozone can be predicted in the future 50 years. Based on the above experiment results, the global ozone change in the future 50 years can be forecasted, namely the atmospheric ozone layer will return to its 1980's standard by the middle of this century at the global scale.