In 2014, China adjusted its "city categorization standard." The newly defined megalopolises and metropolises are under unprecedented pressure from various eco-environmental problems, making them suitable representatives for exploring the state of urban ecosystem health. In this study, we establish a two-layer indicator system to assess the urban ecosystem health and choose 33 indicators grouped into social, economic, transportation, facility, land, and management subsystems, with the aim of correlating human activities with the structure, vigor, resilience, and health of the urban ecosystem. We integrate subjective and objective methods to determine weights at different levels through the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), the analytic hierarchy process, and information entropy. In particular, we develop a spatial TOPSIS technique by introducing a Euclidean-distance-based weight to rank the health of the cities' ecosystem in terms of the spatial effects among these cities. The results reveal that megalopolises such as Beijing, Shanghai, and Guangzhou have superior social and economic subsystems, whereas other megacities have advantages in transportation, facility, land, and management subsystems. From 2005 to 2010, the gaps among these cities in terms of urban ecosystem health significantly reduced regardless of the weight determination method. Not all indicators involved can help realize a better urban ecosystem. Nevertheless, they provide a reference point for making specific regulations to control human activity and improve eco-environmental management.