With the rapid development of Urban Aerial Mobility (UAM), electric Vertical Take-Off and Landing (eVTOL) vehicles are expected to become an important commuting mode for future travelers. Traditional traffic assignment models, which are primarily based on two-dimensional road networks, face significant limitations in representing the three-dimensional path structure and cross-modal transfer behavior introduced by eVTOL services. These limitations hinder accurate modeling of route choice, impedance estimation, and network equilibrium, thereby posing new challenges to conventional traffic assignment frameworks. To address these issues, this study proposes a land–air collaborative network equilibrium model that incorporates Park and Ride (PR) behavior, aiming to comprehensively analyze travelers’ mode choice and route choice in such a hybrid transportation system. Considering the substantial overlap between PR and other travel modes, a Cross-Nested Logit (CNL) model is adopted to capture the similarity structure among modes. In addition, an M/M/1/C queueing model is used to characterize waiting behavior at vertiports, based on which impedance functions for different types of links are formulated. For solution, an iterative framework based on an improved gradient projection (iGP) algorithm is developed, enabling the joint realization of mode split and flow assignment and ultimately achieving user equilibrium. Numerical experiments conducted on the Sioux Falls network verify the effectiveness of the proposed model and algorithm. Results indicate that when PR-based mode choice is considered, more than 30% of travelers are attracted to adopt eVTOL services, which substantially influences network equilibrium and flow distribution. The findings offer new insights into the coordinated planning of UAM and ground transportation systems.