 This paper proposes a new fish detection network, FD-UnderScoreNet, which uses a modified version of the YOLO-V7 algorithm to detect nine different types of fish species in underwater images. The authors found that their modified version of YOLO-V7 outperformed other algorithms in terms of accuracy and speed, achieving a mean average precision, map, of 14.29%, compared to the original YOLO-V7's map of 8.9%. Additionally, the authors tested their FD-UnderScoreNet against other state-of-the-art algorithms, including YOLO-V3, YOLO-V3-TL, YOLO-V3-BL, YOLO-V4, YOLO-V5, faster RCNN, and the latest YOLO-V7 model, and found that their FD-UnderScoreNet outperforms all of them in terms of both accuracy and speed. This article was authored by Hassan Malik, Ahmad Naeem, Shazad Hassan, and others.