Over the past 60 years, land use and land cover (LULC) have changed dramatically within the village landscapes of China’s Yangtze Plain, especially in croplands, causing significant changes in their soil total nitrogen (TN) and total phosphorus (TP). Estimates of long-term changes in the region′s cropland LULC areas and TN and TP in the top 30 cm of soils from the 1940s to 2002 were made using a regional landscape sampling and upscaling approach. Fine-scale LULC area estimates were obtained by field-validated high-resolution ecological mapping of 12 regionally-stratified landscape sample cells based on historical aerial photographs (1942) and IKONOS imagery (2002). Current soil TN and TP was measured at points selected at random within ecologically-distinct landscape features chosen within landscape sample cells using a regional-area-weighted stratified sampling design. 1940s-era soil TN and TP data were obtained from regional historical sources published before 1965. Long-term changes in cropland LULC areas and in soil TN and TP across the village landscapes of the Yangtze Plain were then estimated by integrating these data using a multivariate regional optimization and resampling procedure combined with Monte Carlo uncertainty analysis. From the 1940s to 2002, 47% of village landscape area (86×103 km2) in China′s Yangtze Plain underwent a substantial change in LULC class, of which 33% was cropland transformation. Cropland areas declined by 18.6% (-16.0×103 km2) overall, caused by a 21.5% (-18.5×103 km2) decrease in paddy and a 1.7% (-1.5×103 km2) decrease in rainfed perennial cropland, coupled with net increases in irrigated perennial (3.5%, 3.0×103 km2) and annual (20%, 1.7×103 km2) croplands. However, even with this substantial decline, paddy remains the most extensive land use in the village landscapes of China′s Yangtze Plain. This study demonstrates with 98% probability that soil TN storage in the top 30 cm of croplands across the village landscapes of the Yangtze Plain decreased from the 1940s to 2002, while soil TP storage did not change significantly. Cropland soil TN storage declined by a total of 7.2 Tg N, caused mainly by an 8.0 Tg N decrease in paddy soil TN attributed to the 22% decline in paddy area. Over the same period, soil TN storage in rainfed perennial croplands also decreased by 0.7 Tg, while soil TN storage increased by 1.3 Tg in irrigated perennial and by 0.7 Tg N in irrigated annual croplands. Soil TP storage in croplands did not change significantly across village landscapes, partly because paddy soil TP density variation was very large, and because TP density increased at the same time that paddy area decreased, such that the net 29% median increase observed in soil TP density yielded only a 76% probability of a net increase. Similarly, soil TP storage in rainfed perennial croplands changed very little, decreasing by 0.3 Tg P with a 75% probability of a net decrease. In contrast, soil TP storage increased significantly in irrigated perennial and annual croplands, by 0.7 and 0.4 Tg P, respectively. By combining a regionally-stratified sample of fine-scale landscape features with current and historical soil data in a regionally-weighted upscaling analysis, long-term changes in cropland areas and soil TN and TP storage were revealed across the agricultural areas of China′s densely populated Yangtze Plain.