To address the issues of material damage, complexity, and time consumption in traditional methods of measuring wood elastic modulus, this study proposed a novel prediction method based on XGBoost. This method first analyzed various factors affecting the wood elastic modulus and identifies the key feature indicators. Then, the XGBoost algorithm was selected to construct a prediction model for the wood elastic modulus. The key input parameters of the model were tested and adjusted, and the model’s predictive performance was evaluated by comparing it with actual measurements. Additionally, the gray-box model established based on XGBoost can provide explanations for feature importance to some extent and offer sufficient information to help understand the model predictive behavior. This allows for the adjustment and optimization of model parameters, offering an advantage over black-box models that cannot explain the decision-making process and internal logic. The results showed that the XGBoost algorithm achieved a prediction accuracy of 94.2% for the wood elastic modulus, demonstrating high accuracy and reliability. This validated the effectiveness of the model in predicting wood elastic modulus, overcoming the shortcomings of traditional measurement methods and improving measurement efficiency and accuracy.