Croplands play an important role in China’s agricultural production and food security. However, the shortage of water resource due to climate change and unsuitable utilization poses heavy pressure on agricultural water use in China. Water productivity (WP), defined as the amount of crop production per unit of water consumption by croplands, provides a useful indicator for quantifying where the water can be used more effectively. To date, the spatiotemporal variations of cropland WP in China and its controlling factors at the interannual scale remain poorly understood. In this study, a remote sensing-based ecosystem model (i.e., Breathing Earth System Simulator, BESS) was applied to quantify and analyze the spatiotemporal variations and driving factors of cropland WP in China during 2001–2015. The results showed cropland WP in China had high spatial heterogeneity, ranging from 0.27 to 3.91gC kg⁻¹ H₂O with an average of 1.86 ± 0.30 gC kg⁻¹ H₂O. Dry farmland and paddy field differed considerably in WP values across different regions. During 2001–2015, WP of most croplands (88%) exhibited significantly increasing trends, and dry farmland generally had greater increasing trends than paddy field among all regions. Contribution analysis revealed that the spatiotemporal variations of cropland WP during 2001–2015 were mostly attributed to remarkable increase of crop yield (i.e., GPP), except for some croplands in northwestern regions (e.g., GX) where WP variations were regulated by cropland water consumption (i.e., ET). Furthermore, we examined the driving factors of cropland WP interannual varibility (IAV), and found the dominant factor of WP IAV varied greatly between cropland types and regions. Overall, precipitation was the primary driver of cropland WP IAV at the national level, followed by air temperature and solar radiation. Besides, drought also plays a great role in manipulating WP IAV, especially the medium and long-term drought.