Reasonable forecast of the main cargo throughput has decisive significance for ports to improve their competitiveness and economic benefit. The traditional forecasting methods and models have been difficult to forecast the current needs. In this paper, in order to enhance accuracy of the main cargo throughput, three methods, which are Pearl curve model, moving average method and three exponential smoothing method of time series are used. Based on these three methods, a combined forecasting method is done, and it integrates the advantages of the former three methods, overcomes the uncertainty of a single forecasting method, improves the prediction results and better fits Dalian Port's practical prediction.