This paper examines the long-run relationship between income and urban air pollution using a joint distribution dynamics approach. This approach enables to estimate the transition process and long-run distribution and to examine the mechanisms behind the evolution process. The approach is applied to a unique panel data of CO2, SO2 and PM2.5 (particulate matter smaller than 2.5μm) for 286 Chinese cities over the period 2002-2014. Strong persistence in the transition dynamics suggests that this convergence process may require a long time. The distribution dynamics analyses indicate that multiple equilibria are the major characteristics in the long-run relationship between income and urban air pollution in China, which implies that inter-regional technology spillover may be an important way to accelerate convergence. Our results further support the existence of poverty-environmental trap in PM2.5 concentrations. Thus, new environmental models are expected to be developed to explain this new stylized fact. The findings provide strong support for taking more aggressive measures that consider income and urban environment simultaneously to reduce poverty and air pollutions together in the Chinese cities.