With the improvement of people’s quality of life, filmmaking has become an increasingly important part of the business industry. The score prediction before the movie is released has great reference value for the research of the related recommendation work. Although several studies have been done in this field, it is still necessary to improve the prediction performance of the movie scoring model. In this paper, an effective model is developed to predict movie score by constructing the feature engineering and using XGBoost for prediction. The whole experimental procedure strictly complies with the realworld scenario. The empirical evaluation results demonstrate the outperformance of our approach has good prediction ability on the real data.