As a key node of domestic and international trade, the way to improve the operational efficiency of port has become a hot topic. For the current inefficiency problem of domestic ports, the super-efficiency DEA model to measure the operational efficiency of China''s major coastal ports from 2007 to 2016 is used. It has been shown that the average super-efficiency values of ports are on the rise, but there is a significant decline in individual year. Since 2016 is a pivotal year for the integration of shipping ports, the three-stage super-efficiency DEA method to measure the efficiency of the ports in 2016 is used. The first stage uses the results of the super-efficient DEA model to analyze the second stage. The stochastic frontier model (SFA) is used to incorporate environmental variables such as hinterland GDP that affect port efficiency into the model, eliminating random errors and the impact of external environment on efficiency. The third stage uses the adjusted input-output value to re-evaluate the port efficiency and obtain a more accurate port efficiency ranking. Compared with the first-stage super-efficiency results, the difference between the previous and the following values is significant. Except Qingdao Port and Shanghai Port, all of the ports have not achieved the effectiveness of DEA, and the space for improving is relatively large. Finally, based on the throughput-efficiency matrix, the ports are divided into four types: lean, strong, obese and thin, and targeted improvements are proposed for each type of ports.