<p>Amid growing concerns about urban water pollution and intense human disturbances, urban wetlands face severe ecological pressures. However, existing wetland health assessment studies have predominantly focused on large-scale watersheds, lakes, or coastal zones, with insufficient attention to urban wetland ecosystems. Traditional evaluation models also suffer from drawbacks such as subjective weighting biases and difficulties in processing high-dimensional indicator data, and few studies have systematically quantified the relative importance of key factors in urban wetland health. To address these gaps, this study targeted Chengdu’s Baihetan Urban Wetland Park, establishing a practical ecological health assessment system based on the Driving-Force-Pressure-State-Response (DPSIR) framework and innovatively adopting the Simulated Annealing-Projection Pursuit (SA-PP) coupled model to objectively process high-dimensional data, eliminating subjective judgment biases. Assessment of 2018–2021 monitoring data revealed that Baihetan Wetland Park generally maintained a relatively healthy ecological state, with a steady upward trend in health levels over the four-year period. Weight calculation results based on the optimal projection direction vector of the SA-PP model indicated that ecosystem biological characteristic indicators had the most significant impact on overall wetland health, followed by resource and environmental pressure indicators. This study’s novelty lies in integrating the DPSIR framework with the SA-PP model for urban wetland-specific assessment, filling the methodological gap in urban wetland health evaluation under intense human disturbance. Beyond providing a scientific basis for the targeted management of Baihetan Wetland Park, the findings offer a transferable tool for other urban wetlands and broader insights for coordinating urbanization and ecological protection.</p>