In artificial intelligence and educational data mining, student academic performance prediction is a challenging concern.Therefore, early detection and prevention measures are important in educational institutions. Several studies have been proposed by researchers in the fields of education and artificial intelligence, in which it is challenging to predict the academic performance of the students. Also, it is a major challenge for researchers to identify which machine learning techniques are very accurate in determining a student’s academic success. To solve this challenge, in our study, we explored surveys and studies in this area. These surveys and studies from different countries mainly focus on trying to identify effective predictive models of students’ academic performance prediction. In this study, we try finding answers to study concerns such as what techniques are available to determine the correct sample size and what sampling techniques are accessible in machine learning. This paper enclosed the critical studies based on student academic performance prediction and techniques available for determining accurate sample size and sampling process from 2018 till 2022 with research objectives, research gaps, results, and discussion on findings, and future research recommendations.