Global Positioning System (GPS) is proven to be an accurate positioning sensor. However, there are several sources of errors such as ionosphere and troposphere effects, satellite time errors, errors of orbit data, receivers errors, and errors resulting from multi-path effect which reduce the accuracy of low-cost GPS receivers. These sources of errors also limit the use of single-frequency GPS receivers due to their less accurate data. Therefore, its important to reduce the effect of errors on GPS systems. In order to cope with these errors and enhance GPS systems accuracy, Differential GPS (DGPS) method can be used. The problem with this method is slow updating process of differential corrections. In this paper, three algorithms based on Kalman Filtering (KF) are proposed to predict real-time corrections of DGPS systems. The efficiency of proposed algorithms is verified based on the collected of actual data. The experimental results carried out in field tests assure the high potential of these methods to get accurate positioning data. The results show that KF with variable transition matrix is better than other methods; so that its possible to reduce the Root Mean Square (RMS) of positioning errors in low-cost GPS receivers to less than one meter.