In the transportation sector, fleet management relies heavily on effective route planning and optimization. This procedure entails figuring out the best routes for a fleet of cars to ensure on-time delivery while reducing travel time, fuel consumption, and operating expenses. The intricacy of routing problems, which can involve high-dimensional data with several constraints, presents a challenge. To give transparency and interpretability in decision-making, this research suggests an intelligent route optimization system that combines Explainable Artificial Intelligence (XAI) with the Firefly Algorithm (FA) for feature selection and optimization. By iteratively improving solutions based on the firefly' brightness and appeal, the FA—which was inspired by the natural flashing activity of fireflies—is helpful in tackling complicated optimization issues. The goal of this study is to improve fleet management's route optimization while cutting expenses and increasing efficiency. Deep learning, AI transparency, fleet management, route optimization, and machine learning are among the keywords. When FireflyXRO was compared to more conventional algorithms (Dijkstra's, A*, and Genetic), it showed advances in several important performance areas. Compared to conventional approaches, FireflyXRO avoided congestion in eight more zones, reduced travel time by 18%, and saved 12% on fuel use. While adjusting in real-time-to-real-time traffic data, the algorithm-maintained user satisfaction and interpretability scores of 9.2 and 9.5 out of 10, respectively. These outcomes demonstrate how well FireflyXRO works to improve fleet management route optimization, which raises operational effectiveness and lessens environmental impact.