Due to the influence of many factors such as reservoir lithology and pore structure, it is difficult to identify oil and water layers with traditional complex fluid identification methods, and the identification efficiency is low. Based on the traditional identification methods such as map method, overlap method and logging method, the logging response characteristics can be integrated, and the basic geological data, logging data and oil-water relationship can be fully used to identify the fluid properties. However, the unstable properties and strong heterogeneity of tight sandstone reservoirs, coupled with the complex relationship between pore structure and oil and water, lead to low identification efficiency and strong subjectivity, which limits the identification effect. Taking the Chang 8 tight reservoir in Yanchi area as the research object, this paper proposes a method for fluid identification of tight reservoir based on decision tree. In this method, parameters such as porosity, permeability and oil saturation are calculated as the characteristic values of the decision tree through logging response characteristics and reservoir fluid characteristics of the study area, combined with logging data such as resistivity, sonic time difference, neutron and density, etc. Finally, the Gini index is calculated according to the principle of decision tree. The smaller the Gini index is, the higher the accuracy is, the decision tree model is established. The characteristics of reservoir flow are predicted. The test results show that the accuracy of the decision tree prediction results is 90%, which is basically consistent with the oil test results, which verifies the effectiveness and reliability of this method for tight reservoir prediction. [ABSTRACT FROM AUTHOR]