Crown-of-Thorns Starfish (COTS) and Long-Spined Sea Urchin populations are significant indicators of coral health. COTS, which feed on coral polyps, are maintained at a population of less than 15 per hectare to avoid rapid and extensive destruction of coral reefs. Long-Spined Sea Urchins are macroalgal glazers maintained at a population of five per sqm to maintain low algal cover. The proposed system is a YOLOv8-based model for automated quantification of Crown-of-Thorns Starfish and Long-Spined Sea Urchins from underwater video inputs. This model was trained with a custom dataset of COTS and long-spined sea urchin images for precise detection and integrated with ByteTrack and Supervision for its counting capabilities. Performance metrics results show a mAP score of 0.989 with an F1 score of 0.97 at a 0.59 confidence interval. The model exhibited an accuracy of 86.16counting in 5-ft saltwater. The YOLOv8 model’s performance metrics and testing results show that the model proves to be capable of counting crown-of-thorns starfish and long-spined sea urchins in an actual marine environment.