For a hybrid vehicle, an appropriate energy management strategy is crucial to distribute the energy to satisfy the power performance of the vehicle. The plug-in hybrid electric vehicle (PHEV)’s driving conditions in Fuzhou, China were simulated in this study considering the power components, driver’s habit, and dynamics module using Matrix Laboratory (MATLAB) and Simulink. In the simulation, the adaptive equivalent consumption minimization strategy (ECMS) simulation model was constructed. The equivalent factors for initial states of batteries, driving distances, and driving conditions were optimized using the variable-step Firefly Algorithm with an established library of optimal equivalent factors. A penalty function was introduced using the reference curve to dynamically adjust the equivalence factors in real time. The backpropagation neural network was used to build the Simulink model for driving condition determination, and its prediction accuracy reached 88.7%. Compared with the rule control strategy, adaptive ECMS improved the fuel economy by 7.77 and 4.48%. The results serve as a reference for developing an integration strategy for sensor data to enhance PHEV performance.