Access control (AC) systems are crucial for safeguarding sensitive data and resources, yet the increasing complexity of dynamic environments has underscored the need for more accurate and efficient solutions. Therefore, there is a growing need to enhance the accuracy, efficiency, and decision-making capabilities of AC systems. Machine learning (ML) techniques offer promising solutions to address these challenges and automate access control processes. This paper conducts a systematic review of ML techniques in AC systems, involving a comprehensive analysis covering the identification and classification of ML models and their specific applications. Through a meticulous examination of 62 relevant studies published between 2000 and 2023, the review reveals a predominant focus on innovative solutions and adaptations to enhance access control decision-making. Comparative studies play a significant role in assessing different ML approaches, with a focus on identifying the most effective methods for addressing access control challenges. Notably, Attribute-Based Access Control (ABAC) and Role-Based Access Control (RBAC) emerge as the predominant models, while Support Vector Machine (SVM) tops the list as the most commonly employed ML technique, followed by Random Forest (RF) and Decision Tree (DT). The review evaluates the performance and effectiveness of ML models in AC systems, highlighting their strengths and limitations. Additionally, it addresses significant challenges in the field and identifies potential directions for future research. This systematic review provides valuable insights into the current state of ML-based access control systems, fostering further advancements in this vital domain.