This paper proposes a distributed processing method applicable to GPS receivers in a network to generate a regional ionosphere map accurately and reliably. For accuracy, the proposed method is operated by multiple local Kalman filters and Kriging estimators. Each local Kalman filter is applied to a dual-frequency receiver to estimate the receiver’s differential code bias and vertical ionospheric delays (VIDs) at different ionospheric pierce points. The Kriging estimator selects and combines several VID estimates provided by the local Kalman filters to generate the VID estimate at each ionospheric grid point. For reliability, the proposed method uses receiver fault detectors and satellite fault detectors. Each receiver fault detector compares the VID estimates of the same local area provided by different local Kalman filters. Each satellite fault detector compares the VID estimate of each local area with that projected from the other local areas. Compared with the traditional centralized processing method, the proposed method is advantageous in that it considerably reduces the computational burden of each single Kalman filter and enables flexible fault detection, isolation, and reconfiguration capability. To evaluate the performance of the proposed method, several experiments with field collected measurements were performed.