Abstract: Based on A-F multidimensional poverty measurement method, a multidimensional relative poverty identification index system is constructed to identify the multidimensional relative poverty in Hotian County, a deep poverty-stricken area in southern Xinjiang. It is found that about 52.127% of farmers have multidimensional poverty when κ=3 is taken as the poverty threshold. When decomposing the contribution rate of each dimension to multidimensional relative poverty, it is found that employment status and income level contribute the most to the multidimensional relative poverty of farmers. At κ=3, the multidimensional relative poverty level of women in the study area is significantly higher than that of men. The incidence of relative poverty among the elderly in the study area is relatively common; according to the type of farmers ' poverty, it is found that the poverty level of low-income cardholders is the deepest, and the relative poverty level increases with the increase of dimension. The countermeasures to solve the problem of multidimensional relative poverty are put forward.

Key words: multidimensional poverty, relative poverty, poverty identification, poverty governance, A-F method