Abstract: The stress, deformation and seepage status of Shuibuya high concrete face rockfill dam (CFRD) are estimated according to its monitoring data since completed three years ago. The radial basis function neural network (RBF-ANN) is adopted to get constitutive model parameters based on the measured displacement data. The inversion results of Shuibuya CRFD show that the calculated settlements fit well with the measured ones, and the dam deformation is within a reasonable range and tends to be stable.