本文旨在从统计学的角度对台风行进方向进行统计分析，建立台风方向序列的时序模型。第一部分，先定义台风的方向并给出了计算方法，即根据西北太平洋的热带气旋路径数据集中台风路径上点的经纬度计算出行进方向，以便对此方向数据进行时序分析。第二部分，介绍了时间序列分析的基本概念以及模型。第三部分，给出了对台风方向序列建模的步骤，分析中采用的平稳性检验为ADF检验，白噪声检验为Ljung-Box检验，相应的介绍也在本部分给出。第四部分，本文选取台风弗雷德以及龙王两个著名台风，进行具体的建模分析，并给出了模型中参数的估计。第五部分给出结论：台风的行进方向序列符合带有漂移项的随机游走模型，并对数据集中所有台风的参数估计进行统计分析，发现漂移项常数具有正负不对称性，以及误差方差与纬度之间具有弱相关性。

The purpose of this paper is to make a statistical analysis of the direction of typhoons to establish a time series model of the typhoon direction sequences. In the first part, the definition of the direction of the typhoon and the method to calculate are given, the latitude and longitude of the typhoon track from the path data set of tropical cyclones over the Northwest Pacific can be used to calculate the direction sequences which will be analyzed by time series models. The second part introduces the basic concepts and models of time series analysis. In the third part, the steps of modeling the typhoon direction sequence are given. The stability test used in the analysis is ADF test, and the white noise test is Ljung-Box test. The corresponding introduction is given in this part. In the fourth part, this paper selects two famous typhoons which named Fred and Longwang to build time series model, and gives the estimation of the parameters in the model. The fifth part gives the conclusion that the direction sequences of typhoons follow the random walk model with drift term, and the parameter estimation of all typhoons in the data set is statistically analyzed. It is found that distribution of the drift constants has positive and negative asymmetry, and the error variance has a weak correlation with the latitude.