This paper discusses the setting up of a multivariate statistical method in selecting the useful soil quality indicators for soil quality assessment under agroforestry pattern. The of soil quality has been recognized as a tool to determine the sustainability of land resources, especially in agroforestry development. The study was carried out at Upper Citarum Watershed of Bandung district, west java province, Indonesia. The soil samples were taken with purposive sampling under agroforestry pattern. Principal component analysis (PCA) was used as the multivariate statistical method to identify the minimum data set (MDS); scoring of each indicator, and data integration in the index of soil quality. The MDS consisted of four soil chemical indicators and represented 83.6% of the variability of data, i.e., pH, and exchangeable Calcium (exch Ca), organic Carbon (org C), and exchangeable Natrium (exch Na) respectively. The soil quality index (SQI) was categorized under agroforestry pattern as moderate. The artificial agroforestry-based coffee with an intercropping system (timber woods, multi purpose trees and horticultures) provides better soil quality.