The frequency of urban floods is increasing due to extreme precipitation events, necessitating ur-gent improvements in urban flood resilience (UFR). To effectively assess the level of UFR, this contribution proposes an improved cloud model UFR assessment model based on pressure-state -response (PSR) model and game theory. A UFR index system is constructed based on the PSR model, incorporating urban constituent elements, resilience characteristics and resilience realiza-tion process. Based on game theory, the analytic hierarchical process (AHP) and entropy weight method (EWM) are combined to build an improved cloud assessment model of UFR. Taking Zhengzhou City as an example, the assessment results show that with nine years of continuous ef-forts, the UFR of Zhengzhou has improved year by year from 2012 to 2020 and reached medium resilience level III from lower resilience level II. The pressure layer is found to be risky, the state layer is moderate resilience, and the response layer is more responsive. The assessment model is feasible, and the results are consistent with the actual situation. Optimization suggestions are made in terms of infrastructure and ecological environment, early warning and emergency man-agement, and flood resilience knowledge learning, respectively, to provide theoretical support and decision-making reference for the improvement of the UFR management level.