The acquisition of plant height data mainly relies on manual measurement, which is not only time-consuming and labor-intensive, but also subjective. In order to solve this problem, a plant height measurement system was constructed, by controlling the rotation of the two-dimensional LiDAR, the three-dimensional reproduction of the plant is realized, and the height information is obtained. Using this system, for single plants such as Green dill and Evergreen, use the side overall contour scanning method to obtain the side projection and calculate the height; for the Boxwood Hedgerow and Golden Leaf Ligustrum Hedge, the random sampling consensus algorithm (RANSAC) is used to fit. The algorithm is revised, and then the curve detection-normal nuclear density detection method is used to filter out the outliers in the canopy, and the height is calculated. Experimental results show that the maximum relative deviation between the plant height obtained by this system and the manually measured value is less than 6%. This system has good practical value for plant height measurement.