Abstract: The validation of large-area land cover data is a costly, labor-intensive and time-consuming task with the participation of several experts from different regions and different institutions. Pre-viously, people always used single-machine or off-line software to realize it because of lack of methods and tools which could support online and collaborative validation. Focused on resolving this problem, this paper involves the concept of "Internet⁺" and technologies of service computing to land cover data validation. Based on this ideal, an "Internet⁺" oriented service model for land cover data validation is proposed, and then a new online and collaborative validation process is designed. Additionally, to fit the automatic and efficient computing requirements, the LSI(landscape shape index)-based sampling algorithm is put forwarded. At last, we introduce the development of GLCVal platform and an international cooperation practice for validation of GlobeLand30 based on it. The practice is supported by GEO(Group of Earth Observation) and UN-GGIM(the United Nations Initiative on Global Geospatial Information Management), and is promoted enormously by GLCVal platform in the realization of cross-regional cooperation and active participation of experts from different countries. Practice results show the model and algorithm proposed by this paper can provide new idea and method for high resolution land cover data validation in large region, and also increase efficiency of validation.