In the context of the digital transformation of education, traditional learning models based on standardized approaches demonstrate limited effectiveness and fail to account for individual learner characteristics. In this regard, the development of adaptive curricula based on educational data analysis and the application of Data Mining methods has become increasingly relevant. The aim of this study is to develop and experimentally validate a model for adaptive curriculum design that ensures the personalization of the educational process and enhances its effectiveness. The methodological framework integrates approaches from Learning Analytics, machine learning, and data mining. The study employs classification, clustering, regression, and association rule mining techniques. The empirical basis of the research includes data from 2,500 students, collected from Learning Management Systems (LMS), encompassing demographic, academic, behavioral, and log data. The experimental results demonstrate that the Random Forest model provides the highest prediction accuracy for academic performance (Accuracy = 0.89). Cluster analysis identified three groups of learners: high-performing students, average-performing students, and at-risk learners. Association rule mining revealed a significant relationship between student activity levels and learning outcomes. Comparative analysis indicates that the implementation of the adaptive learning model leads to an increase in the average academic score by 12 points and an improvement in the success rate by 18.0%. The findings confirm the effectiveness of a data-driven approach and its importance in improving educational quality. Thus, the study demonstrates the high potential of integrating Data Mining methods into educational systems and substantiates the need to transition toward adaptive and intelligent learning models.