As farmed fish are exceedingly sensitive to changes in parameters such as hazardous toxins, pH, temperature, ammonia, and gas presence, water quality control is important in aquaculture. The water quality needs to be regularly monitored and regulated if fish are to remain healthy and to maintain a quality production. The catla fish will get lethargic and eventually goes into a coma and die if the ammonia level rises high enough. Predicting the ammonia levels in aquaculture is therefore crucial. By gathering information from the water’s humidity, pH, ammonia levels, etc. This can serve as the data foundation for an early warning system and better catla fish farm management. Deep learning model CNN-LSTM technique, has been used to forecast ammonia in aquaculture. The LSTM model’s cells make it effective at identifying long-range dependencies, which eliminates the variable gradient problem. Additionally, CNN is typically a more effective and precise method of handling categorization issues.