Worldwide socioeconomic development has resulted in huge irretrievable environmental problems in various ecosystems. This study employed seven coastal watersheds in two provinces, Zhejiang and Fujian, China forming a gradient to testify the environmental Kuznets curve (EKC) interactions between socioeconomic development and environmental impact at a watershed scale. Annual socioeconomic indicators, including gross domestic product (GDP) and its components, registered population (agricultural and non-agricultural population), and electricity consumption, and annual discharges of chemical oxygen demand (COD) and ammonium were collected at a county level, and land use pattern to generate watershed level dataset in the period of 2011–2016. Results indicated that non-agricultural GDP per capita of the non-agricultural population and discharge of COD or ammonium per unit of total GDP were top-ranked pair-indicators significantly fitting the EKC model instead of the classic GDP per capita and pollutants. The development of seven selected watersheds have passed the turning point of the EKC and entered impact-reducing development stages along the EKC, i.e., the three Zhejiang watersheds are at the low-impact development stage, the Huotong Stream watershed from Fujian province was at impact-declining development stage right, and other three Fujian watersheds were at medium-impact development stage. In term of the environmental impact indicator, pollutant discharge per unit of total GDP serves as a development impact indictor per se. These findings might provide an EKC-based approach to support and strategize the watershed management for sustainable development in the world.