Many people go by subway in China. Huge passenger flow brings much trouble to the police and passengers, such as crowded carriages, long waiting times and low-efficiency transport. So it is important to know the traffic jam before it brings many problems. With the data from Tianchi Competition, this paper analyzes and predicts the subway ridership of a station in Hangzhou based on time series and linear regression. Taking line 3, station 5 as an example, there were two peaks which have many passengers in the station. Combined with the results, it proposed that traffic police need to pay more attention to the rush hours. People who do not need to commute can avoid these times. The government may also use this prediction results in the subway service management, as well as in the planning for the future development and subway lines projection.