This paper introduces an improved underwater trash detection model, YOLOv8-C2f-Faster-EMA, based on the YOLOv8 algorithm. The model is designed to enhance the detection of small-scale underwater debris, addressing the challenges of high miss and false detection rates in aquatic environments. The YOLOv8-C2f-Faster-EMA algorithm optimizes the backbone, neck layer, and C2f module for underwater characteristics and incorporates an effective attention mechanism. This optimization improves the accuracy of underwater litter detection while simplifying the computational model. Empirical evidence shows that the proposed method outperforms the conventional YOLOv8n framework, achieving a 6.7% increase in precision (P), a 4.1% surge in recall (R), and a 5% enhancement in mean average precision (mAP). The method's superior performance in detecting small targets and its potential for integration into remote sensing ventures make it a significant advancement in marine conservation and environmental monitoring. The paper also discusses the architecture of the YOLOv8 network, the FasterNet architecture, and the EMA module, detailing their contributions to the overall performance of the model. Experimental results using the Trash_ICRA19 dataset demonstrate the model's effectiveness in detecting various underwater targets, including plastic debris, biological matter, and ROVs, with high precision and recall.This paper introduces an improved underwater trash detection model, YOLOv8-C2f-Faster-EMA, based on the YOLOv8 algorithm. The model is designed to enhance the detection of small-scale underwater debris, addressing the challenges of high miss and false detection rates in aquatic environments. The YOLOv8-C2f-Faster-EMA algorithm optimizes the backbone, neck layer, and C2f module for underwater characteristics and incorporates an effective attention mechanism. This optimization improves the accuracy of underwater litter detection while simplifying the computational model. Empirical evidence shows that the proposed method outperforms the conventional YOLOv8n framework, achieving a 6.7% increase in precision (P), a 4.1% surge in recall (R), and a 5% enhancement in mean average precision (mAP). The method's superior performance in detecting small targets and its potential for integration into remote sensing ventures make it a significant advancement in marine conservation and environmental monitoring. The paper also discusses the architecture of the YOLOv8 network, the FasterNet architecture, and the EMA module, detailing their contributions to the overall performance of the model. Experimental results using the Trash_ICRA19 dataset demonstrate the model's effectiveness in detecting various underwater targets, including plastic debris, biological matter, and ROVs, with high precision and recall.