Abstract: Increasing human activities bring remarkable pressure to island ecosystems, which are ecologically vulnerable. Comprehensive and accurate island ecological evaluation is the basis of island protection and management, however, it puts forward high requirements for fieldwork data. The island field investigation is difficult and costly due to the obvious spatial isolation and complex terrain conditions. The rapid development of remote sensing technology has provided abundant data source for the island ecological evaluation, and it is urgent to explore whether the island ecosystem can be accurately and comprehensively evaluated based only on remote sensing data. In the present work, 18 inhabited islands in Changhai County, Liaoning Province of China were selected to demonstrate such an study on island ecological evaluation based only on remote sensing to reduce the cost and provide a feasible method. A total of 10 types and 27 sub-types of land covers were systematically depicted and the ecological information of remote sensing was utilized to establish the island remote sensing ecological index (island remote sensing ecological index, IRSEI), which contains three components, namely, landscape, vegetation, and soil. The indices were calculated at 30×30 m grid scale and island scale, and their correlations with the geographical parameters were analyzed at the grid and island scales. The results indicated that altitude, slope, and distance to coastline have positive correlations with IRSEI at the grid scale, and altitude and slope have higher contributions than distance to coastline; at the island scale, distance to mainland has a positive correlation with IRSEI. This study has validated the comprehensiveness and accuracy of IRSEI for assessing island ecosystem and its spatial exhibitions at different scales, and the index can be applied to similar studies on island ecosystems in different regions and at multiple spatial and temporal scales.