In recent years, the sector has witnessed a vast shift in the direction of the adoption of Artificial Intelligence (AI) across various fields because of its fantastic potential to research huge-scale facts, identify styles, and help accurate and efficient decision-making. AI plays a vital position in enhancing overall performance, saving time and sources, and enhancing carrier fine—particularly in important sectors like education. AI algorithms vary in kind, including traditional device mastering methods including long -term memory (LSTM), multilayer Perceptron (MLP) and Conversional Neural Networks (CNN), in predicting the overall performance students. The aim of this study is to analyze the application of different deep study techniques to predict the students' performance based on the old data collected from the Department of computer science at the Faculty of Information Technology. A methodology was implemented for collecting and ‎preprocessing student data, followed by training the models ‎on the prepared data. The ‎model ‎obtained was tested the use of check information amassed from the same ‎domain. Experimental effects showed that CNN achieved the ‎maximum F1 Score factor of 0.9032 to improve both LSTM and MLP. ‎Despite the minor adjustments, all models validated robust ‎accuracy (87.5%), confirming their reliability in instructional ‎data analysis. Additionally, a comparison was conducted among deep neural networks to highlight the differences in their performance. These findings emphasize the significant role of artificial intelligence and neural networks in advancing the educational process by enabling early intervention strategies and improving learning outcomes............... Keywords: ...............Artificial Intelligence in Education, Neural Networks Student Performance Prediction, Deep Learning, LSTM, MLP, CNN.