<p>This paper addresses the correlation between vehicle battery state of charge (SOC) and driver pedal behavior by investigating the coupling relationship between vehicle SOC and driving style recognition. It proposes a driving style recognition method based on the density-optimized WOA-K-means++ (DWOA-K-means++) algorithm—considering SOC influence. First, 6 months of real-world new energy vehicle data were preprocessed and segmented into multiple driving data intervals based on SOC (0–100%), enabling SOC-based segmented analysis. Next, principal component analysis was employed to derive comprehensive feature parameters characterizing driving styles under varying SOC influences. Subsequently, the DWOA-K-means++ clustering algorithm was applied to classify driving styles based on these parameters. Finally, an artificial neural network was employed to validate driving style recognition. The results indicate that drivers with different driving styles exhibit varying sensitivities to vehicle SOC under different SOC conditions. The method proposed in this paper achieves high accuracy in driving style recognition, laying the foundation for further optimization of energy management strategies for new energy vehicles.</p>