<p>The regional bus services operated by the Karnataka State Road Transport Corporation (KSRTC) are essential for linking urban and rural area. However, its face persistent operational challenges related to route planning, scheduling inefficiencies, and fluctuating passenger demand. To address these issues, this study proposes a transparent hybrid machine learning fuzzy logic framework for assessing passenger satisfaction and supporting data-driven service improvements. Primary passenger survey data are analyzed using Principal Component Analysis (PCA) to reduce correlated service-quality attributes into five key components most prominently Comfort, Punctuality, and Booking Convenience that combine explain 94.1% of the total variance. The Analytical Hierarchy Process (AHP) is employed to assign relative importance weights to critical service parameters, ensuring a structured and consistent decision framework. A Decision Tree (DT) classifier is then developed using categorical variables, numerical features, and PCA-derived scores to predict passenger satisfaction levels. The PCA-DT model demonstrates robust performance, achieving a test accuracy of 94.56% and a stable fivefold cross-validation accuracy of 94.0% ± 1.3%. Fuzzy logic is applied as an independent post-hoc interpretive layer to capture nuanced passenger perceptions and validate model outputs while avoiding label leakage. Statistical analysis reveals a strong association between service quality and satisfaction outcomes (χ² <i>(4)</i> = 45.2, p &lt; 0.<i>001)</i>. Feature importance results highlight punctuality and comfort as the dominant determinants of satisfaction. Overall, the proposed framework provides an interpretable, auditable, and operationally relevant tool for real-time passenger satisfaction assessment in public transport systems.</p>