The rural sustainability is one of the most important issues in rain-fed agricultural areas; however, because of multiple obvious and latent driving factors, a few models have been suggested covering the fluctuation of sustainability in these sensitive areas. Therefore, this study takes advantage of a combination of multivariate and multi-criteria decision making to analyze the association of rural resources with sustainability in northwest of Iran. The entire demographical, agricultural and natural standard data of 143 villages (21 variables) were weighted and ranked by Shannon–Winner index and simple additive weighting (SAW) method, respectively. The synthesized dataset was grouped into two categories of highly sustainable (ZSAW > 0) and low sustainable (ZSAW < 0) villages. Factor analysis (FA) was separately performed on the grouped data to identify possible interrelationships among the variables. The results of FA illustrated a significant difference between low and highly sustainable villages with respect to type and contribution of the resources in sustainability. In both cases, the first five rotated factors, including FA1 (human), FA2 (agronomical), FA3 (mechanization), FA4 (rangelands) and FA5 (wheat yield) vectors, accounted for about 99% of the variation. At low sustainable villages, the two most important driving factors were agronomical (51.1%) and rangelands (17.3%) which explained 68.9% of the variation; however, for highly sustainable villages the main factors shift to the human resources (22.1%) and agronomical (43.9%) with 66% contribution. These findings show that reliance on natural resources is a sign of unsustainability in the rain-fed agro-ecosystems, whereas, with increasing human resource contribution, villages become more sustainable. The study also confirmed that human resources are significantly and positively correlated with sustainability index (ZSAW) in highly scored villages. Furthermore, among the human resources, population density
