Abstract: In this paper, using the K-Means quick clustering method, sequence association rules, Bayesian network, QUEST decision tree, C&R decision tree, CHAD decision tree, support vector machine (SVM) and neural network data mining techniques, such as the QUEST for Xi'an urban residents sports consumption household survey data for empirical analysis.The result shows that food, clothing, shelter and line etc.Necessity consumer spending is still the main spending of Xi'an urban residents.As a development and consumption of enjoyment in Xi'an urban residents, sports consumption expenditure of daily consumer spending is still in a low level, the proportion of food and clothing live line, education, saving and other consumer spending and sports consumption expenditure significantly correlation between characteristic;Xi'an urban residents sports consumption of the project have the characteristics of diversity, is still in the by material object to participating sports consumption and sports consumption and ornamental type transition stage of the sports consumption, buy sports clothing, shoes, caps, sports fitness equipment purchase and accept the professional fitness instructor obviously correlation between characteristic;Xi'an sports consumption levels of urban residents in different population characteristics, sports psychological trends and lifestyle variables show the different characteristic.