The standardization of river system structures has reduced hydrological connectivity, exerting significant pressure on both river ecology and urban development. The precise diagnosis and proposal of optimization strategies remain key challenges. This study utilizes graph theory to delineate the river network topology in Chengdu city and establishes an evaluation index system (α, β, and γ) for hydrological connectivity. Spatiotemporal changes in Chengdu's hydrological connectivity from 2000 to 2020 are analyzed, alongside a validated optimization scheme focusing on “pressure dispersion” among river network nodes. The findings reveal a simplification of Chengdu's river network structure over the past two decades, marked by a reduction of 72 nodes and 81 river segments. Particularly in the city center (region II), these changes have significantly diminished overall connectivity. Following the implementation of the “pressure dispersion” optimization method, the distribution of critical nodes in Region II has been more equitably, enhancing resilience against disasters like floods by diversifying the burden on key nodes. Furthermore, the hydrological connectivity indices in Region II showed improvements post-optimization, with the α, β, and γ indices rising by 75%, 7.26%, and 7.26%, respectively, thereby validating the efficacy of the approach. Analyzing variations in hydrological connectivity parameters across river systems identifies vulnerable links and fundamental issues, laying groundwork for future improvements. The proposed optimization methodology offers insights applicable to cities worldwide grappling with similar connectivity challenges.