The ionosphere crucially impacts on Global Navigation Satellite System (GNSS) positioning accuracy and integrity. Recently some network-based methods have shown the potential to construct a regional/global vertical total electron content (VTEC) or grid ionospheric vertical delay (GIVD) map for accuracy augmentation purposes. However, how to use these advanced methods for integrity augmentation has not been adequately investigated. The authors have investigated a regional ionospheric integrity monitoring strategy based on the radial basis function neural network (RBF-NN), using GNSS TEC observations. Similar to the SBAS approach, the GIVD map is constructed so as to enhance positioning accuracy, and the corresponding grid ionospheric vertical error (GIVE) map is constructed for protection level calculation to enhance positioning integrity. To reduce the GIVD residuals and the GIVE values, the local ionospheric spatial activity index (LISAI) is proposed as an indicator of local ionospheric spatial activity level. The RBF-NN structure parameters are able to be adaptively determined via hierarchical clustering. Modeling results in the China region have verified that the proposed GIVD modeling method is slightly better than the classical WAAS-Kriging method. The proposed GIVE modeling method significantly outperforms WAAS-Kriging, achieving an improvement of around 46% and 25% during the ionospheric calm and active periods, respectively.