Fast wavefront reconstruction is crucial for improving the temporal frequency of adaptive optics systems, in which a mass of subapertures is used. In this paper, we present a novel block zonal reconstruction algorithm based on Southwell geometry to speed up the wavefront reconstruction from Shack-Hartmann wavefront sensor measurements. Therein, we use the theory of computational complexity to install a novel optimal block zonal strategy to get the best size of the subwavefront and give the verification through simulations. Compared with the classical Southwell entire wavefront reconstruction algorithm, the algorithm based on the optimal block zonal strategy needs only a few milliseconds to finish the reconstruction process from 100×100 subapertures. Moreover, we analyze the superiority to use our algorithm to realize the phase reconstruction of the unconnected subwavefront, which cannot be reconstructed via traditional methods. Further, the simulation and experiments show that the precision of the algorithm based on the proposed optimal block zonal strategy is comparable with the commercial wavefront sensor HASO, the time consumption is much less than that of the traditional zonal reconstruction, and it is applicable to the square wavefront, circular wavefront, and local unconnected wavefront. Our proposed algorithm can be widely utilized in astronomical observation, laser transmission, and remote sensing.