Establishing an accurate ship fuel consumption (SFC) prediction model is crucial for achieving ship energy efficiency optimization (SEEO), and enhancing the green development for the maritime sector. A GA-CNN-LSTM (GCL)-based prediction method of energy usage for ships adopting wind-assisted propulsion system (WAPS) based on the operational data is constructed. Firstly, input features for the prediction model are selected by analyzing the characteristics of ship operational data. Then, a GCL-based SFC prediction model is established. The research results show that the proposed model’s prediction accuracy is as high as 0.9706 in terms of R2, which shows better performance than other models, including SVR, BP, CNN, GRU, LSTM, and CNN-LSTM. Finally, the energy usage monitoring and analysis system for ships adopting WAPS is developed based on the established SFC prediction model, which is an important platform for SEEO and control of ships adopting WAPS. The proposed SFC prediction model can not only provide scientific basis for improving the ship energy efficiency, but also contribute to promoting green development of the maritime sector.