The high cost of real-time kinematic (RTK) differential GPS units required for autonomous guidance of agricultural machinery has limited their use in practical auto-guided systems especially applicable to small-sized farming conditions. A laser range finder (LRF) scanner system with a pan-tilt unit (PTU) has the ability to create a 3D profile of objects with a high level of accuracy by scanning their surroundings in a fan shape based on the time-of-flight measurement principle. This paper describes the development of a LRF-based autonomous navigation algorithm for a head-feeding rice combine harvester that could automatically follow straight rice rows based on real-time detection of rice uncut edges. A motor-driven crawler type platform operated on a myRIO real-time controller was constructed to develop a steering control algorithm suitable for such a tracked type-driving mechanism. Noise data existing in raw dataset were removed to extract rice row profiles without unpredictable disturbances by using the revised random sample consensus (RANSAC) method. Boundary points between uncut and cut edges were then determined using the maximum method. The 3D points defined in terms of the LRF sensor coordinates were converted into the vehicle coordinates by considering the platform movement and PTU rotation in order to create a 3D field map of uncut edges for autonomous harvesting. A PID steering control algorithm based on a linear relationship between the lateral deviation and heading error of the mobile platform was implemented. Laboratory tests showed that the PTU operation improved the ability of the uncut edge detection algorithm to detect the target in the presence of interfering objects as compared to that measured without use of the PTU, showing a decrease in lateral RMSE from 21.8 to 5.7 cm whereas there was little change in heading RMSE < 2 deg. A fundamental navigation experiment showed that the mobile platform-mounted LRF scanner system could guide the motor-driven platform following straight and curved edge lines of artificial targets with an acceptable level of oscillation at a traveling velocity of 0.14 m/s. Therefore, the use of a laser based real-time path generation and tracking algorithm would be  feasible in automatically guiding the rice combine harvester.