Objective A medical image can reflect the physiological and pathological information of the body tissues in both structural and functional aspects. In this paper, three-dimensional reconstruction of the heart and cardiovascular system is performed on the basis of a series of medical images to assist surgeons. In this paper, region growing method is used in a three-dimensional model, and three sub-methods of it are compared in the segmentation effect of the cardiovascular system model. Methods Combined with the software package of VTK (Visualization Toolkit) and ITK (Insight Segmentation and Registration Toolkit), this paper firstly uses ray casting method to implement three-dimensional reconstruction of a series of CT scanning images to generate a model of thoracic cavity; subsequently utilizes three forms (confidence connecting, neighborhood connecting and connected threshold) of 3D region growing method to segment the images, in order to obtain the simulation result of the cardiovascular system model. Results It is possible to use 3D-reconstruction technology to realize the reconstruction procedure of tissues in a human chest described by the cardioangiography sequences. Due to the comparison of the three segmentation forms, neighborhood connecting contributes a rough model which loses much more information about the anatomical structure, besides, the vessel branches are less and the vessel diameter is smaller. Conclusions By using VTK and ITK, cardiovascular system model can be constructed via the technology of volume rendering and image segmentation under the guidance of ray casting method and 3D region growing method; the connected threshold algorithm and confidence connecting algorithm preserve more details.