Anomaly detection systems play a significant role in maintaining the health and productivity of aquaculture ecosystems. In this project, we propose a novel approach utilizing machine learning techniques to analyze both fish behavior and the physical aspects of the aquaculture environment. Our system leverages pre-collected datasets obtained from sensors that monitor critical environmental parameters such as temperature, pH levels, dissolved oxygen, nitrate concentration, and fecal coliform. Its purpose is to find anomalies that may indicate threats against the balance or permanence of an aquacultural setup. Our system monitors fish behavior by analyzing their speed and sudden directional changes. Fish moving erratically or at unusually high speeds often signal distress or discomfort, potentially caused by environmental stress, disease, or poor water quality. This paper presents an analysis of a few machine learning models to detect anomalies considering both the environmental and behavioral aspects so that preventive action can be taken to reduce risk at the earliest and save lives among fish populations.