An important part in farmland quality evaluation is summarizing the county-level results into provincial result. As the foundation of the national summary, the summary accuracy directly affects the distribution of the national farmland land quality. Because of the summary base map lags behind, a method was proposed for quickly updating summary base map. Meanwhile, the multiple attribute constraint and nearest-neighbour provincial summary method for farmland quality database was proposed, then the summary software was developed for automating provincial summary. Daxing district of Beijing was taken as an example, then the rationality of summary base map process was analyzed and the distribution of farmland land quality before and after summarizing was compared by the standard distance, mean center and standard deviation ellipse indexes. Results showed that except grade 16, the standard distance reduced after summarizing, which meant that the summary result was more gathered than the county level and the process of summary base map was reasonable. Furthermore, by comparing the mean centre and the directional distribution before and after summarizing, the distribution of farmland quality almost had no change, which indicated that the spatial distributions of the province and county farmland quality were in highly consistent.