This paper investigates the integration of explainable artificial intelligence (XAI) principles into augmented reality (AR) systems for assembly guidance tasks. We identify three key stakeholders—end-users, educators, and system developers—who require different levels of explanation transparency in AR assembly systems. Our explainable AR application, developed using Unity and Vuforia for wind-up car assembly, incorporates decision transparency, confidence visualization, and interactive explanations. Validation with four students demonstrates significant improvements in user understanding, assembly accuracy, and system trust compared to traditional paper manuals. The AR system completed assembly tasks in 2 minutes and 38.85 seconds with comprehensive step-by-step analytics. Results show that explainable AR addresses critical gaps in assembly guidance systems by providing transparent decision-making processes that enhance user engagement and trust across diverse stakeholder groups.