Natural rubber is an important and strategic raw material, used in tires, gloves, and insulating products, that is mainly obtained by cutting the bark of rubber trees. However, the complex contour curve of the rubber tree trunk is hard to fit using a tapping machine. Thus, a trunk contour curve collection would be useful for the development of tapping machines. In this study, an acquisition system based on laser-ranging technology was proposed to collect the point cloud data of rubber tree trunks, and a B-spline fitting program was compiled in Matrix Laboratory (MATLAB) to extract the trunks’ contour curves. The acquisition system is composed of power, a controller, a driver, a laser range finder, and data transmission modules. An automatic extraction experiment on the contour curves of rubber tree trunks was carried out to verify the feasibility and accuracy of using the acquisition system. The results showed that the degree of rubber tree trunk characteristic recognition reached 94.67%, which means that the successful extraction of the rubber tree trunk contour curves and the B-spline fitting program are suitable for the extraction of irregular curves of rubber tree trunks. The coefficient of variation of repeated collection was 0.04%, which indicates that changes in relative positions and acquisition directions have little influence on the extraction and the accuracy of the acquisition system, which are high and stable. Therefore, it was unnecessary to adjust the position of the acquisition device before the collecting process, which helped to improve the efficiency of acquisition considerably. The acquisition system proposed in this study is meaningful to the practical production and application of agroforestry and can not only improve the precision of the rubber tapping process by combining with an automatic rubber tapping machine but can also provide technical support for the prediction of rubber wood volume and the development of ring-cutting equipment for other fruit trees.