Abstract: Through analyzing the data of poverty alleviation in Xiaoba village，Fudong Township ,L.an-cang County,this paper presents a small data analysis method suitable for targeted poverty alleviation. It improves the correlation analysis and reasoning methods of machine learning and data theory , and based on the small data set analysis method of Naive Bayes theory，the accuracy index of poverty all eviation effect is given by this method，which can analyze and evaluate many cases and compare with the evalua-tion results of other evaluation methods. The results show that the method will not be limited by the small number of samples，and it is easy to calculate. It has the comparability , universality and practicability for targeted poverty alleviation.

Key words: ｓｍａｌｌ ｄａｔａ, ｂｉｇ ｄａｔａ, ｔａｒｇｅｔｅｄ ｐｏｖｅｒｔｙ ａｌｌｅｖｉａｔｉｏｎ