Abstract: To address the dual pressure on transportation and power networks caused by the large-scale integration of Electric Vehicles （EVs）, and the limitations of existing charging station planning that neglects user spatio-temporal stochasticity, bounded rationality, and the techno-economic potential of Vehicle-to-Grid （V2G） technology, this paper proposes a location and capacity planning strategy considering user microscopic behavioral characteristics and V2G collaborative potential. A refined spatio-temporal behavior model for vehicles is constructed based on travel chain theory and probability distribution. The Weber-Fechner Law is introduced to quantify user psychological responsiveness to charging prices, time, and queuing. Furthermore, a comprehensive objective function is established, encompassing construction and maintenance costs, generalized user losses, and human-factor associated costs. An improved Memetic Algorithm （MA） integrated with a "greedy pruning" strategy is employed to solve the model.Numerical simulations demonstrate that the proposed strategy reduces the annualized total system cost by 3.8% （approximately 345,000 RMB） through V2G collaborative scheduling while ensuring user satisfaction. Specifically, the direct economic benefit from V2G reverse discharge reaches 114,000 RMB. Compared to traditional genetic algorithms that ignore V2G coordination and user psychological behavior, the optimization accuracy of the improved MA is enhanced by 10.0%. The planning strategy, which integrates user psychological behavior and V2G coordination, effectively avoids facility redundancy and improves the system's economic efficiency and regulation capability. It provides a decision-making reference for the interactive integration of transportation and power networks.