The ship construction period is long, the material cost proportion is large, and it is easily influenced by many factors such as commodity price index and exchange rate, resulting in a large error between the actual completion cost and the quotation estimation. The influencing factors of material cost are identified with Grey Correlation Analysis (GCA) method, a rolling forecasting model of ship material cost is constructed based on Long Short-Term Memory (LSTM) model. The test analysis is conducted with the material cost data of 53 64 000 t bulk carriers in 63 months from a shipbuilding enterprise and the corresponding influencing factors. The results show that the error between the forecasting data and the actual data is within the acceptable range, which can prove the validity of the selected method and the constructed model. The research results are of practical significance for the real-time cost forecasting and control of ship construction process.