In this paper, we aim to employ the least squares support vector regression (LS-SVR) for the spatio-temporal modeling of the ionospheric total electron content (TEC). This model utilizes simple linear equations to solve the system of equations, thereby reducing the computational complexity and enhancing both the speed of convergence and the accuracy of the results. We utilized observations from 15 GPS stations in north-western Iran from day 193 to day 228 in 2012. The results of the LS-SVR model were compared with those of support vector regression (SVR), artificial neural networks (ANN), adaptive neuro-fuzzy inference system (ANFIS), Kriging model, global ionospheric maps (GIM), and the International Reference Ionosphere 2016 (IRI2016) as well as TEC values obtained from GPS. The accuracy of all models was evaluated and interpreted at interior and exterior control stations. The analyses indicate that the average root mean square error (RMSE) for the ANN, ANFIS, SVR, LS-SVR, Kriging, GIM, and IRI2016 models at two interior control stations are 3.91, 2.73, 1.27, 1.04, 2.70, 3.02, and 6.93 TECU, respectively. Furthermore, the average relative errors of these models at the same control stations were calculated as 15.98%, 9.39%, 7.85%, 6.09%, 11.60%, 12.54%, and 26.56%, respectively. Analysis of the precise point positioning (PPP) method demonstrated an improvement of 50 mm in the coordinate components using the LS-SVR model. The results of this study demonstrate that the LS-SVR model can serve as a viable alternative to global and empirical models of the ionosphere in the studied area. The LS-SVR model provides a high-precision local ionosphere model.