In this work, we propose the computational algorithm to compute the temperature-independent specific surface area of atomistic models of porous carbons. Concisely, the method is to divide the simulation space into fine grid points, place the selected probe molecule on each grid point and then check whether that insertion is accessible or inaccessible regions based on molecular force field. Three models of porous carbon structures are chosen as examples for estimating internal and external specific surface areas and trend of surface curvature. By comparing the computational cost, our proposed technique significantly requires less time-consuming than the physisorption. Therefore, for atomistic models of porous carbons, we recommend that our proposed method be more efficient and accurate than the experimentally and computationally traditional physisorption.