Artificial intelligence plays a significant role in sports analytics, enabling the prediction and optimization of athletic performance based on physiological, psychological, and genetic parameters. This research focuses on developing a predictive model for analyzing athlete performance, specifically estimating Power-w/kg, using the Random Forest Regressor. The dataset consists of 35 athlete records, which were expanded to 70 records through data augmentation. A structured machine learning pipeline was implemented, covering data preprocessing, feature engineering, augmentation, and model training. The model was evaluated using standard performance metrics such as MAE, MSE, RMSE, R2, and MAPE. The results demonstrated strong predictive performance, achieving an R2 score of 0.94 and a MAPE of 5.33%, indicating high accuracy. The study highlights the importance of feature selection and augmentation in improving model generalization. Additionally, the visual analysis of residuals and feature importance provided further insights into model behavior. This research confirms that machine learning algorithms can effectively predict athletic performance, offering a trustworthy and explainable AI-based system. Future work should focus on expanding the dataset, optimizing hyperparameters, and exploring deep learning approaches to enhance predictive accuracy.