Suspicious activity detection, often called unusual event detection, is an essential component of video analytics that plays a crucial role in identifying abnormal occurrences. While this process faces challenges such as human variability, changes in lighting, and variations in shapes, advancements in machine learning and deep learning techniques are meant for effective solutions. This survey aims to constructively review the current research on suspicious human activity detection, providing insights into the latest developments in machine learning and deep learning methodologies and highlighting state-of-the-art datasets. It will also identify key challenges in detecting video anomalies based on human behavior and explore the effectiveness of deep learning approaches. Additionally, this review showcases its significance in enhancing surveillance and security efforts.