Severe weather such as heavy precipitation is of great harm to the normal operation of civil aviation. The accurate prediction of precipitation is helpful to the safe and stable operation of civil aviation and other enterprises. By preprocessing the precipitation time series data collected by automatic meteorological observation of airport runway system(AWOS), training and testing sample sets are provided for deep learning. Then, long short-term memory model (LSTM) and time series convolution network (TCN) models are constructed respectively to realize the prediction of precipitation in the next 1 to3 hours. Then, the prediction accuracy of the two models is compared and analyzed. The results show that the prediction performance of TCN model is better than that of LSTM model. The R2 of TCN model is 0.96, 0.91 and 0.86 for the forecast of precipitation in the next 1 to 3 hours, respectively.