High-resolution ionospheric structure estimation with low latency is essential for the stable maintenance of various future services. However, no papers have yet addressed this issue. In this paper, a new algorithm for real-time high-resolution ionospheric tomography within a large area is developed. The algorithm uses the property that the ionospheric electron number density does not change significantly in 30 seconds, the observation interval of the Total Electron Content (TEC), to estimate the unknown distribution using the known immediately preceding distribution as the initial solution. The conjugate gradient method is used to search for the solution, but it is converged very quickly by using a diagonal scaling matrix, which is a very simple preprocessing step. Using the algorithm on actual ionospheric data, we found that it is possible to estimate the electron number density distribution with an unprecedentedly high resolution of 0.25° square for TEC data within a 1-minute interval. The algorithm also reconstructed the ionospheric structure during the occurrence of Medium Scale Traveling Ionospheric Disturbances (MSTID), and was able to estimate the detailed its horizontal and vertical structures. This indicates that this algorithm is useful for real-time estimation of the three-dimensional structure of ionosphere. Further development of a system utilizing this algorithm will lead to further studies of the ionospheric mechanisms.