The measurement method based on a single uncalibrated image is a new research hotspot in recent years, which mainly uses the structure information contained in the image itself to realize the geometry information extraction of spatial objects. However, the previous studies always carried out measurement in a single plane. If we want to achieve the measurement of multiple planes, each plane needs to be carried out separately, so we need more known conditions, which is difficult to meet in many cases. Therefore, the original method needs to be further developed. In this paper, an algorithm model based on the single image measurement is proposed, which can use the known segments information in different planes, so as to achieve the multiple planes measurement using fewer known conditions. This algorithm reduces the quantity requirement of known conditions, and improves the original single image measurement model. The information required in the algorithm is usually easy to be obtained in structured scenes, which is useful for the user's on-demand measurement. In addition, the measurement can be repeated many times, thereby reducing the workload of field measurement. From experiments, the results show that the relative errors of indoor and outdoor distance measurement are almost both better than 3%.