Mechanical weeding is a necessary means to produce organic soybean in vast area and is the most effective for seedling of weed. However, the growth of weed is inconsistent; we must rely on the quantity of closely weeding to achieve the expected results. Organic cultivation adopts strip planting to reduce the cost of weeding, and uses the cultivator to remove weeds. This operation is slow and with high cost, and the artificial weeding is necessary in the later stage. In contrast, the field weeding robot can provide mechanical or physical prevention of weeds via monitor and image analysis. Therefore, this study aims to effectively increase production and solve the problems of labor shortage and high cost on labors by combining an unmanned ground vehicle with a developing smart laser weeding system for autonomous precise weeding. An intelligent image recognition is adopted to identify the position of the weed, and then through the laser tissue fibrosis to achieve the goal of precise weeding.