The invention discloses a three-dimensional modeling method of an object based on a depth sensor, comprising the following steps: S1, a data acquisition step of reading data from the depth sensor to obtain three-dimensional point cloud data; S2, a preprocessing step of separating the point cloud of the object from the whole scene, removing abnormal points, extracting the boundary of the point cloud of the object, and obtaining the point cloud of the object under each visual angle; S3, a point cloud registration step of registering an object point cloud under an adjacent perspective to obtain aconversion matrix between the object point clouds under the adjacent perspective; S4, a global alignment step of fusing the point clouds under each visual angle into the global point clouds by usinga conversion matrix to obtain a three-dimensional point cloud model; S5, meshing and texture mapping step, converting the three-dimensional point cloud model into a mesh model, and then mapping the color information to the surface of the mesh model to obtain a three-dimensional texture model. On the basis of guaranteeing the modeling accuracy, the method improves the modeling speed, and has high use and popularization value.