The Relief Degree of Land Surface (RDLS) represents the undulation of a region's terrain and is determined by the elevation change relative to the flat land. RDLS is a macroscopic indicator of topographic features that are closely linked to local climate, vegetation, biodiversity, land use/cover, and human activities. These factors significantly influence ecosystem patterns, processes, and functions, which are fundamental to the provision of ecosystem services. Therefore, revealing RDLS's effects has great implications for improving ecosystem services. Previous studies have examined the distribution pattern, value, and temporal and spatial evolution of ecosystem services based on regional topographic gradients. However, the response process of ecosystem services and their trade-off/synergistic relationship with RDLS in the watershed remains unclear. To address this gap, we conducted a study focused on the Jiuquxi watershed, which is the core watershed in Wuyishan National Park. And we explored the response of key ecosystem services, ecosystem service bundles and their trade-off/synergistic relationships to the RDLS in Jiuquxi watershed from 2009 to 2016, using relief degree of land surface model, InVEST model, K-mean clustering analysis, ecological services change index (ESCI), and ecosystem services trade-off degree (ESTD). The results showed that (1) the ecosystem services including biodiversity, carbon sequestration, water yield and water purification increased with the increase of RDLS during 2009-2016, as did their gain area. (2) In the Jiuquxi watershed, ecosystem services can be categorized into three types of bundles:eco-fragile bundles, where all four services are low; eco-regulating bundles, where all four services are high with carbon sequestration being dominant; and eco-conservation bundles, where all four services are high with biodiversity being dominant. The eco-fragile bundles clustered in the regions with low RDLS value and eco-regulating bundles in the area with high RDLS value. While the eco-conservation bundles had no typically spatial pattern for ESCI. From 2009 to 2016, 19% of the areas in eco-conservation bundles in low RDLS region were transformed into eco-fragile bundles and eco-conservation bundles, and 18% in high RDLS region. (3) The pattern of ecosystem service trade-offs/synergies in response to RDLS was identified:an inverted U-shape curve with secondary or quaternary RDLS at its peak (e.g., the trade-off between carbon sequestration and biodiversity, and the synergistic between carbon sequestration and water yield) and the N-type curve with secondary and quaternary RDLS as the inflection point (e.g., synergistic between water purification, carbon sequestration and water yield, and the trade-off between water purification and biodiversity).