When simulating the lubrication characteristics of sliding bearings,the lubrication characteristics simulation based on existing CFD is computationally expensive and time-consuming,and are not suitable for multiple iterations in the optimization design of oil film.Polynomial structure selection technology was proposed to construct the lubrication characteristic surrogate model of wind power sliding bearing.In which,a genetic intelligent distribution technology is proposed to reduce the number of CFD simulations with genetic samples,based on the error reduction ratio,the effective items that have a great influence on the lubrication characteristics are screened,the invalid items with little influence are eliminated,so the number of items is reduced to improve the efficiency and accuracy of the polynomial surrogate model of the lubrication characteristics of wind power sliding bearing.The lubrication characteristic surrogate model realizes the polynomial fitting of the strong nonlinear mapping relationship between lubrication parameters and lubrication characteristics,and the number of samples required to be collected is less than that of the traditional radial basis function surrogate model construction method,which provides a new research idea and technical approach for complex nonlinear modeling.