The paper provides insights into the application of machine learning in evaluating academic performance, the importance of the study is highlighted in adopting the right decision-making through prediction and analysis of ideas that work to improve the performance of faculty members and evaluate performance and other matters in a behavioral manner, as the algorithms were extracted and incorporated with machine learning equations to extract the best variables by comparing the performance of several models in classifying cases affecting the quality of education and academic performance of the teacher, as the worst was excluded through the KNN K-nearest neighbors algorithm where the value of K was 3. This helps the algorithm in classification and regression problems by integrating repeated queries for machine learning data inputs and predicting descriptive and behavioral academic performance. The study dealt with classifying the sample variables into groups of variables, which are (professional development, self-training, evaluation and feedback, improving the design of educational curricula, how to manage time, activating automated administrative tasks, and strategies to facilitate cooperation), the samples included 256 lecturers, technical and administrative academic at the Northern Technical University, the opinions of 256 valid questionnaires from selected samples were sent statistical analysis, consisting of 33 data entries for two variables that were distributed in the study community and the study applied the method of statistical inference and prediction with finding the values of linear regression and the relationships of influence and correlation for the study variables.