Abstract: Mineral deposits are often deformed after mineralization, which is, however, less concerned in the current three−dimensional (3D) prospectivity modeling. This paper selected the Dayingezhuang structural altered rock type gold deposit as a case study and used a structural restoration method based on triangular irregular network (TIN) to reconstruct 3D orebody and ore−controlling fault, analyzed and compared the mineralization structure and ore−controlling factors before and after restoration and finally completed the 3D mineral prospectivity at depth. The results show that the structural restoration method can eliminate the variation of spatial distance and dip angle of fault and orebody caused by deformation. The reconstructed mineralization distribution has a stronger spatial autocorrelation feature that is shown as the change of scattered mineralization distribution to spatially continuous at the offset parts. In addition, the reconstructed prediction model has higher performance than that without restoration under the same parameters, indicating that the correlation between the mineralization distribution and ore control factors is more significant. Therefore, the three−dimensional metallogenic prediction modeling with integration of structural reconstruction has improved the propectivity accuracy and can provide a reliable reference for deep prospecting in the Dayingezhuang deposit.