The strong heterogeneity and complex connectivity of carbonate reservoirs in the Moxi area of the Sichuan Basin result in significant errors in conventional logging prediction of permeability. In order to improve the accuracy of carbonate rock permeability prediction, a permeability prediction method based on digital core is proposed. This article takes the carbonate reservoir of the Dengying Formation in the Moxi area of the Sichuan Basin as an example, and constructs a three-dimensional digital core model based on core CT scanning data. Using connectivity algorithms to obtain connected pores, dividing each pore and throat through separation algorithms, and obtaining the geometric parameters of each pore and throat. After obtaining the geometric parameters of pores and throats, they are divided into two spatial types based on industry standards and existing research results: pores and cracks. The pore structure parameters and crack structure parameters are calculated separately. A comparative analysis was conducted on the correlation between different pore structure parameters and weighted pore structure parameters with permeability. Four pore structure parameters sensitive to permeability were selected, including average pore size, average pore throat ratio, average crack opening, and crack complexity. A multiple regression permeability prediction model was established. This article multiplies the pore structure parameters with the corresponding proportion of pore space to obtain weighted pore structure parameters, effectively improving the correlation between pore structure parameters and permeability. Compared with the traditional dual medium porosity and permeability model, this model reduces the average relative error of predicting permeability from 266.7% to 30.27%, and improves accuracy by approximately 236.4%. The results indicate that the permeability prediction model for carbonate reservoirs established by this method can effectively improve the accuracy of permeability in fractured and vuggy carbonate reservoirs, providing a new approach for calculating the permeability of complex carbonate reservoirs.