Abstract: Chengdu City has undergone rapid urbanization in recent years, leading to significant changes in its landscape patterns. Based on GlobeLand30 land use and coverage data in 2000, 2010, and 2020, this study determined the optimal analysis granularity using the coefficient of variation, granularity response curves, and an information loss model while also defining the optimal extent using the grid method and a semi-variogram function. Subsequently, a landscape ecological risk assessment model was constructed on the optimal scale, followed by the analysis of both the spatiotemporal evolution characteristics and spatial correlation of landscape ecological risks. The results indicate that the optimal granularity and extent for landscape pattern analysis of Chengdu are 90 m and 5 km× 5 km, respectively. From 2000 to 2020, approximately 0.6% of areas in Chengdu underwent land type conversion annually, with the most significant conversion occurring between cultivated land and artificial surfaces. The landscape ecological risks in Chengdu differed significantly in space, being lower in the west and higher in the east. Low-risk areas were concentrated in the central urban district, expanding outward over time. Among the statistically significant spatial clusters, low-low and high-high clusters were identified as the primary landscape ecological risk types in Chengdu. The landscape ecological risks in the city trended downward overall. However, these risks remained unbalanced between the eastern and western parts, with grassland and cultivated land still facing relatively high ecological risks. Overall, this study can provide a decision-making basis for land use management, pattern optimization, and risk management and control in areas undergoing rapid urbanization.