Abstract: Complex meteorological sea conditions directly affect the safety of ship navigation, and the accuracy of the prediction of offshore wind speed, as a major factor in meteorological sea conditions, is of great significance to the navigation safety and trajectory planning. In order to effectively improve the accuracy of offshore wind speed prediction and overcome the limitations of a single prediction model, the offshore wind form data of Lianyungang station is used as an example study, and the Adaboost algorithm is used to integrate the advantages of multi-models to construct a combined prediction model of offshore wind speed. Four time series prediction models, including BP neural network, GA BPNN, long and short-term memory network and WOA-SVR, are used for wind speed prediction. Considering the prediction effect of a single model, Adaboost algorithm is applied to integrate the GA-BPNN model and WOA-SVR model to construct the combined offshore wind speed prediction model, and the integration accuracy is compared with that of Bagging algorithm. The results show that the root mean square error of the combined prediction model with the Adaboost algorithm is reduced by about 13% and the mean absolute error is reduced by about 16% compared with the single model, which effectively verifies the superiority of the combined prediction model in the prediction of offshore wind speed data, and it is of great significance for the enhancement of navigational safety and the optimization of the trajectory design.