This paper discusses time-varying causalities between climate policy uncertainty (CPU) and green finance market, and filling the research gap in revealing their internal dynamic connections. Using rolling-window technique, the empirical findings demonstrate that CPU exert positive and negative impact on green finance market, represented by the green bond index (GBI) during certain time intervals, which do not entirely support theoretical basis. Similarly, the study also uncovers inconsistent conclusions regarding the influence from GBI to CPU among different causal periods, indicating the function of green finance in alleviating climate change is complex. These deviations between variables can be attributed to factors, such as early-stage green bond market, shifts in climate policies, the impact of COVID-19, and underlying macroeconomic fundamentals. The study identifies heterogeneities of the causal relationships in different time intervals, and the bidirectional influence under specific circumstances, offering fresh insights for existing theoretical framework. Several policy implications are thus provided from multiple aspects, such as standardization of green bonds, innovation of green financial instruments, and development of long-term climate governance strategies. These measures can facilitate a positive role for climate policy and green finance in pursuing harmonious development.