The mixed pixel problem impacts significantly the accuracy of land cover land use (LCLU) mapping, particularly in heterogeneous scenes. Subpixel mapping (SPM) addresses this by predicting LCLU distributions at a finer spatial resolution than the input image. However, as an ill-posed problem, SPM faces inevitable uncertainties of prediction. Existing SPM methods commonly rely on auxiliary data at the target fine spatial resolution, which are generally scarce. To address this issue, this paper proposes a novel SPM approach called Coarser-SPM, which integrates information from images that are coarser than the original input coarse image. These coarser images, can include those with spatial resolutions that follow a non-pyramid (i.e., intersecting) relationship with the original coarse image, can provide complementary spatial information and serve as additional constraints for SPM. To integrate the information from both the original coarse image and the coarser images, Coarser-SPM conducts a joint optimization strategy that optimizes class probabilities at the subpixel scale. Experimental results show that with the integration of the coarser images, Coarser-SPM produces more accurate SPM results, revealing more satisfactory details of the LCLU distributions compared to conventional SPM methods. Moreover, the accuracy of subpixels in the intersection area between the coarser and original coarse scales increases significantly, highlighting the effectiveness of using coarser images to enhance SPM.