Sri Lanka is an important hub connecting Asia-Africa-Europe maritime routes, with abundant rainfall but uneven spatiotemporal distribution, and obvious seasonal water shortage. Monitoring water area change in its inland lakes and reservoirs plays an important role in guiding the development and utilization of water resources. In this study, a rapid extraction model of surface water based on Google Earth Engine remote sensing cloud computing platform was constructed. On the basis of evaluating the optimal spectral water index method, the spatiotemporal variation of typical reservoir and inland lakes and reservoirs in Sri Lanka were analyzed. The results showed that AWEIsh can accurately identify the water boundary, with an overall accuracy of 99.14%, which was suitable for the extraction of surface water in Sri Lanka. The area of Maduru Oya Reservoir showed an overall increasing trend on the basis of small fluctuations from 1988 to 2018, and the monthly area of reservoir fluctuated greatly in 2017, thus water resources management in dry zone should pay more attention to seasonal regulation and control. From 1995 to 2015, the number and area of lakes and reservoirs in Sri Lanka increased in different degrees, mainly concentrated in arid provinces such as Northern Province, North Central Province, Western Province, etc., and the amount of surface water resources increased as a whole.