为解决相对贫困的深层次问题，探讨我国农户多维相对贫困的动态演变及其影响因素，以服务精准扶贫目标。利用Python爬虫和文本挖掘梳理研究热点，结合中国家庭追踪调查（CHARLS）数据，采用A-F多维贫困测度法测算多维贫困指数，并通过随机森林和Logistic回归模型分析关键影响因素。结果显示，我国农户多维相对贫困程度较高，教育和收入维度贡献显著，家庭医疗支出负担是主要致贫因素，因病致贫和因灾致贫尤为突出。政策建议包括：加强教育与技能培训，完善医疗保障体系，强化防灾减灾能力，降低多维贫困风险。

In order to solve the deep-rooted problems of relative poverty, we explore the dynamic evolution of multidimensional relative poverty and its influencing factors of China's farm households, so as to serve the goal of precise poverty alleviation. Using Python crawler and text mining to sort out the research hotspots, combining with the data of China Household Tracking Survey (CHARLS), the multidimensional poverty index is measured by A-F multidimensional poverty measure, and the key influencing factors are analyzed by Random Forest and Logistic regression models. The results show that the multidimensional relative poverty of China's farm households is high, with significant contributions from the education and income dimensions, the burden of household medical expenditures is the main poverty-causing factor, and poverty due to illness and disaster is particularly prominent. Policy recommendations include: strengthening education and skills training, improving the medical security system, strengthening disaster prevention and mitigation capacity, and reducing the risk of multidimensional poverty.