License plate recognition (LPR) systems have attracted the community to expand its benefit and usefulness. It gives easiness to keep tracks of the vehicles that enter and exit the same place. This success is influencing the people to develop mobile LPR systems. In the mobile LPR system, the camera is installed on a vehicle such as the police officer's car. In such systems, the problem is more challenging compared to fixed LPR systems. The detection process becomes crucial in the mobile LPR systems, because the license plate can be captured in various angles. The environment also cannot be controlled, and it affects the performance of LPR systems significantly. Thus, this paper proposed license plate localization based on statistical features. The proposed method achieved promising accuracy 88.46% tested on collected database.