The increase in marine litter is slowly becoming a significant problem, for which various recognition techniques have been proposed and are still being. Artificial Intelligence (AI) based methodologies have emerged as a promising tool to address this challenge. However, adopting AI in marine litter search and monitoring requires high performance and accuracy, interpretability, and explainability, which are essential for building trust in the decision-making process. Explainable AI (XAI) is an emerging research area that aims to make AI models transparent and interpretable, enabling human experts to understand and trust the model’s decisions. In this context, XAI can play a crucial role in improving the effectiveness and efficiency of marine litter search and monitoring by providing insight into the model’s decision-making process and identifying areas for improvement. This paper aims to evaluate using a pre-processing methodology for removing water from underwater image interoperability on a slot attention-based classifier for explainable image recognition using a dataset based on marine debris for searching underwater litter. Experimental results show that the application of the above-mentioned pre-processing technique brings about a significant improvement in underwater image classification.