The mango planting industry in China has been developing rapidly, but the spatial distribution information of mango groves has traditionally relied on slow and inefficient manual surveys. Therefore, it is urgent to study the remote sensing extraction method for mango groves. While numerous studies have focused on forest or agricultural information extraction using remote sensing techniques, studies specifically targeting mango grove extraction are scarce. In this study, we analyzed the suitability of remote sensing images with varying resolutions for the extraction of mango groves. Considering the spacing characteristics of mango groves, we generated remote sensing images with resolutions of 2, 4, and 8 m by downsampling 1 m GaoFen-2 data. The support vector machine method was employed to extract mango groves from these images with different resolutions. The results indicated that the highest accuracy was achieved using remote sensing images with a resolution of 2 m, whereas the extraction accuracy of images with a resolution of 1 m was slightly lower. In contrast, the extraction accuracy obtained from remote sensing images with resolutions of 4 and 8 m was low, as it was prone to misclassify mango groves with cultivated land at these resolutions. Therefore, satellite images with a resolution of 2 m are recommended for the extraction of mango groves, as they offer an optimal balance between classification accuracy and processing efficiency.