<p>With the increase in adoption of deep learning in security applications, the lack of transparency in decision making remains a significant concern. Traditional deep learning models in UAVs make it difficult for security professionals to understand trust and interpret their predictions. This challenge is particularly critical in security, where clarity is essential for detecting threats, preventing attacks, and ensuring compliance with regulatory requirements. To address this issue, our study contributes to the growing field of XAI in UAV security by demonstrating its effectiveness in improving the reliability and transparency of deep learning-based security systems. We proposed ADUX that explores Explainable AI for Deep Learning-based security by integrating explainability techniques into deep learning models used for threat detection and anomaly identification. The proposed approach employs Shapley Additive Explanations to provide human-understandable explanations for model decisions. Experimental results show that incorporating XAI techniques improves comprehensibility with increased trust and enables better debugging, while maintaining model performance comparable to state-of-the-art methods.</p>