The need for satellite navigation technology, with its capability to provide real time positioning using Global Navigation Satellite System (GNSS) has increased tremendously over the years. However, the positioning accuracy tends to degrade as the baseline increases. The degradation comes from different sources of errors, amongst which is the ionospheric error in the signal propagation. During geomagnetic storm, ionospheric scintillations can rapidly vary amplitude and phase of GPS signal or result to loss of lock on a satellite and if enough satellites are affected it can result to loss of positioning services. During the field tests, GPS measurements were taken when the kp-index was at the highest. Two stations observations were used. The first is the observations collected from the vicinity of a base GPS station, while the second observation is from the user’s GPS receiver. With the aid of radio link, data from the user station was sent to an independent computer which is connected to the base station. Both observations from the base GPS station and the user’s station were processed simultaneously implementing double differencing method and the LAMBDA method to correct the integer ambiguities in the carrier phase, the tests were processed using ionospheric weight technique. The formulated algorithms and data were processed using Matlab. With the use of stochastic model and an implementation of an extended Kalman filter bad data were accounted for and eliminated. The end result yields a millimetre level positioning.