 The study develops a novel method to improve cropland data accuracy for the belt and road, B&R, region by fusing and correcting four cropland products using spatial analysis techniques, and the Google Earth Engine, the GE platform with random forest RF algorithm. The corrected product indicates that cropland accounts for 14.94 percent of the B&R region, which is closer to FAO statistics than any individual land cover product. The method exhibits better fitting characteristics, and produces an overall accuracy of 77.54 percent in inconsistent areas. This study has potential applications to update global cropland products by combining training samples with multi-source, remote sensing data from the GE platform. This article was authored by Kuei Li and Erky Shu.