To address the issues of low delivery efficiency and high operational costs in electric delivery trucks due to range limitations and static charging planning in dynamic demand environments, a full-process optimization framework of “initial path-dynamic adjustment-flexible adaptation” is constructed, with real-time path optimization and flexible charging mechanisms as the core. Firstly, a mathematical model is established with the objective of minimizing the total operational cost (including path, charging, time window penalty, and dispatching costs). Secondly, the traditional static charging mode with fixed stations is broken, and the dynamic charging decision-making logic is designed using adaptive large neighborhood search (ALNS). During the dynamic path adjustment stage, the optimal charging stations are selected in real time, and the charging duration is flexibly set for subsequent delivery tasks based on the remaining battery power of the vehicles, achieving deep collaboration between the charging mechanism and the path planning. The simulation results of the case study show that this optimization mechanism significantly outperforms traditional methods in terms of reducing operational costs and improving the utilization efficiency of charging resources, providing a theoretical basis and practical solution for the efficient operation of electric trucks in urban distribution.