Abstract: To accurately predict the fuel consumption of in service ships,analyze the complex and variable influencing factors of fuel consumption,and quantify their respective impacts,this study selects tankers and bulk carriers for operational data collection and preprocessing. A fuel consumption prediction model based on the Extreme Gradient Boosting（ XGBoost） algorithm is established,and factor importance is evaluated using the Gain method. The results demonstrate that the proposed model achieves strong computational and predictive performance,with mean absolute percentage errors of4. 88% and 3. 92% for the tanker and bulk carrier models,respectively. Among internal factors,ship speed shows the greatest influence,with weights of 0. 671 and 0. 429 for the two vessel types. Regarding external factors,navigation environment conditions such as wind and waves also exhibit significant impacts.