The ionospheric electron density provides valuable information as to the ongoing physical processes in this part of the atmosphere and is important parameter to cancel the ionospheric effect in positioning computation. Empirical models such as IRI just provide coarse forecasts of the ionospheric electron density values. Nowadays, GPS and other GNSS systems has become a common tool for analyzing the Earth′s atmosphere. Spherical harmonic expansions are used for global modeling of the ionospheric electron density. Spherical harmonics are global support functions. Furthermore, their efficient application requires regularly distributed data on the globe. In this paper, regional four-dimensional electron density is estimated from IRI and Iranian Permanent GPS Network (IPGN) data in order to improve the empirical models accuracy. The electron density is decomposed into reference and correction parts. The reference part is taken from the IR-I2007 model. Due to the localizing feature of B-spline functions, Euclidean quadratic B-splines and tensor-products is used to model the correction term with respect to longitude, latitude and time. EOFs are used to account for the vertical gradient of the electron density in the correction. Ionosonde direct measurements of ionospheric electron density are used for validating the modeling result.