This article describes a system development for autonomous driving of mini-tram and its experimental results on outdoor autonomous driving. The mini-tram was provided by KRRI (Korea Railroad Research Institute). For the pose estimation, we fused an IMU and a GPS sensor using extended Kalman filter. We chose an NVIDIA Drive PX2 and LiDAR to recognize the outdoor environment. In order to verify the feasibility of the developed system, we performed outdoor autonomous driving experiments.