1. For threatened marine megafauna, such as the Indo-Pacific humpback dolphin (Sousa chinensis), sound conservation planning should aim to ensure demographic and ecological persistence of populations. One method to address this challenge is through ecosystem-based conservation planning (ECP) based on threatened marine megafauna distribution, biodiversity richness, and ecosystem functionality. 2. ECP exercises for the Indo-Pacific humpback dolphin in Chinese and adjacent northern Vietnamese waters were conducted based on the Indo-Pacific humpback dolphin distribution, Hill-2 biodiversity index, and marine net primary productivity. Habitat protection priorities were scored using the program MARXAN and were used to identify special areas for conservation (SACs) for the Indo-Pacific humpback dolphin. 3. To ensure the minimal risk of local extinction, the SACs enclosed a total of 26,858 km 2 of waters, accounting for 49.9% of core habitats of the Indo-Pacific humpback dolphin and 24.3% of the IUCN Red List species in the study region. To enclose 30% of biodiversity richness and ecosystem functionality, a total of 40,179 km 2 SACs was required. 4. In the context of threatened marine megafauna conservation, megafauna distribution can be used to highlight focal areas, while information on biodiversity richness and ecosystem functionality should be factored in. Using surrogates with different ecological niches would be a better tactic to ensure sufficient protection coverage by minimizing the omission bias. 5. The conservation of threatened marine megafauna in coastal and estuarine waters shares similar objectives with global marine biodiversity conservation. The challenges to conduct ECP exercises come from data scarcity and poor data quality in representing distributions of biodiversity features. A certified openaccess database that shares survey effort, occurrence, population density, and habitat mapping at national and local scales is recommended. Such a result requires national planning, investment, and policing. Shiang-Lin Huang and Haiping Wu contributed equally to this article.