Food provisioning service (FPS) represented by grain yield is strongly affected by slope and soil and water conservation (SWC) measures in mountainous region. Current FPS assess method mainly depend on regional census statistical data, and it didn’t consideration local variability in slope and SWC measures, which may lead to errors. Based on field monitor, this study qualified the relationship between slopes, SWC measures and grain yield per hectare (GYPH), and then proposed a data down-scaling method (DDM) in estimations of FPS in a mountainous region. It showed that corn yield per hectare (CYPH) was negatively related to slope; a 1-degree increase in slope resulted in a 267 kg∙ha-1 decrease in CYPH on farmland without SWC measurers (FWOM) and a 336 kg∙ha-1 decrease in farmland with SWC measures (FWM). Although FWOM accounted for 52.47% of the total farmland in the study area, it only provided 29.86% of FPS. FWM < 15° was the key area for providing FPS, which indicated that the traditional FPS assessment method (using the census statistical data) was inaccurate compared with DDM. DDM can help to improve the accuracy of ecosystem services assessment and trade-off analyses, and thus contribute to more targeted land use planning in mountainous region.