搜集2008-2014年新疆巴音郭楞蒙古自治州(简称巴州)梅毒月发病数据,采用时间序列分解方法(Time Series Decomposition methods)探讨该地区梅毒月发病率的季节性,建立ARIMA(0,1,1)(1,0,1)_(12)时间序列模型,模型预测值的动态趋势与实际发病率基本吻合,平均绝对百分比误差MAPE=17.36,具有较高的预测精度,可以较好的预测短期内梅毒的变化趋势,为梅毒的预防控制措施提供可靠依据. To collect monthly incidence data of syphilis in Bayingolin Mongolian Autonomous Prefecture of Xinjiang Uygur Autonomous Region from 2008 to 2014, the seasonal incidence of syphilis in the area was investigated by time series decomposition method (Time Series Decomposition methods), ARIMA (0,1 , 1) (1, 0, 1) _ (12) time series model, the dynamic trend of model predictive value is basically consistent with the actual incidence, the average absolute percentage error MAPE = 17.36, with higher prediction accuracy, Predict the trend of syphilis in the short term and provide a reliable basis for the prevention and control of syphilis.