Ionospheric delay is recognized as a major error source for satellite navigation systems. The ICAO SBAS (satellite-based augmentation system), defined by SARPs (international standards and recommended practices) documents, has a capability to make a correction to ionospheric delay effects on GNSS in addition to satellite clock and orbit. It broadcasts users the vertical ionospheric delays in meters at the grid points (IGP; ionospheric grid point) located at every five by five degrees in latitude and longitude. The SBAS is the international standard system for global and seamless satellite-based navigation, so it would be used even in the equatorial or low latitude regions, while in fact the ionosphere has the significant activities in such regions. The equatorial anomalies affect on the large-scale structure of electron density of ionosphere which perhaps might be difficult to be corrected by SBAS ionospheric correction messages. SBAS providers who serve to the magnetic equator and the low magnetic latitude regions need to investigate ionospheric effects in such regions in terms of SBAS ionospheric error correction. Actually Japan has been operating its own SBAS system called MSAS (MTSAT-based satellite augmentation system). It is covering wide range of latitude and the lowest magnetic latitude in the coverage is nearly ten degrees magnetic north, where is very close to the equatorial region. Especially for predicting and improving actual performance of MSAS, we have been investigating the ionospheric effects around Japan. According to the analysis for MSAS shows that some ionospheric correction has a large uncertainty due to improper modeling of the ionosphere. A reason for this is the less number of observations around the IGP, occurring at the edge of the service area, which induces large protection levels and lowers availability. Another reason for improper modeling is that the region of data collection to be input to fitting process is too wide in comparison with the spatial scale of the ionosphere to be fitted. The current algorithm, called planar fit, is designed for the planar ionosphere so the radius for data collection is set as large as possible; However, at the southern islands of Japan there is the equatorial anomaly of ionosphere the ionosphere cannot be modeled as a plane. The problem is that there is no process to make an estimation again with some adjustment of ionosphere model in case of an invalidity of ionosphere model. A trip of the storm detector means an invalidity of ionosphere model used, but does not mean invalidity of other ionosphere models; Even in such a case, we can try fitting process again with an adjusted parameters for the ionosphere model. Here we propose shrinking the fitting radius as a adjustment of model parameters if the storm detector trips. At each IGP, in case that storm detector trips and shows an invalidity of the ionosphere model, we can (i) Shrink the fitting radius and reduce an IPP; or (ii) Shrink an IPP at the southernmost, i.e., the IPP nearest to the equatorial anomaly. According to the simulation, using (i) or (ii) methods, the uncertainty of ionosphere estimation is reduced and the availability of SBAS navigation improves. In the paper, the authors discuss the problem to be solved, the basic idea of the proposed method, and the effects of the method to reduce the uncertainty of ionospheric estimation. Then practical results simulated with the actual data including both ionospheric quiet days and stormy days are shown. The proposed method would contribute to the performance improvement of the SBAS at the equatorial and low latitude regions.