At present, tight oil and gas reservoirs must adopt fracturing technology to obtain productivity, which will not only transform the reservoir, but also bring reservoir damage. Taking Chang-7 member of Ordos Basin as the research object, the relationship between physical properties of tight oil reservoir and fracturing fluid damage is analyzed based on experimental analysis of reservoir physical properties, cast thin sections, electron microscope scanning, X-ray diffraction and sensitivity test. Using the traditional damage evaluation method requires a large number of cores, and core resources, as a nonrenewable precious resource, have been paid more and more attention. Therefore, the use of prediction is conducive to protecting core resources, reducing experimental costs, and improving work efficiency. Therefore, a mathematical prediction model of RBF neural network is proposed, which establishes the complex nonlinear relationship between the physical properties of Chang 7 reservoir and fracturing fluid damage in Ordos Basin. Taking 22 groups of data of Chang 7 reservoir as training data, the fitting rate of training data is 90%. Taking the other two groups of data as detection data, the error between prediction and actual experiment is less than 10%. The prediction shows that the error inside and outside the sample predicted by RBF neural network is small, the prediction accuracy of the model is high, the generalization ability is strong, and the prediction value is closer to the value obtained by laboratory experiments than BP neural network, which can provide a good theoretical basis for fracturing fluid optimization.