Traditional inspection methods are difficult to achieve the accurate inspection of the quailty of prefabricated buildings. To solve the problem above, this paper proposes an automatic extraction method of geometric dimensions of prefabricated components of buildings based on point cloud model. The paper conducts this method under the following steps: firstly, the typical prefabricated component point cloud is separated from the overall building point cloud by taking advantage of the spatial location relationship and geometric characteristics of the building structure. Meanwhile, feature extraction of component point cloud is used to measure dimension information and the perpendicularity of the wall is detected by calculating the included angle between the fitting plane normal vector and the vertical direction unit vector. The perpendicularity of the column is detected by extracting the central axis of the column from the point cloud slice. Moreover, the flatness of the board is represented by the mean square error from each point to the fitting plane. The research takes the point cloud data collected from the construction site of a prefabricated building as an example, removes and segments their outliers, after which it can manage the automatic extraction and measurement of point cloud data features. The measurement accuracy, above 95%, verifies the feasibility and effectiveness of the method, which provides a reference for automatic segmentation, dimension measurement and parametric management of typical prefabricated components.