In recent years, optical fiber shape sensing technology has been widely studied in various fields, and has been widely used in robot, medical, aerospace, industrial equipment structure monitoring and submarine cables. With the change of application scenarios and the gradual improvement of measurement performance requirements, the research needs of optical fiber shape sensing technology are becoming increasingly urgent. At present, the research on fiber shape sensing is mainly divided into two directions. One is the shape sensing technology based on FBG, which takes advantage of the wavelength drift of FBG under strain and realizes shape measurement by writing FBG on multi-core fiber, which has the advantages of high precision and simple data processing. In this direction, some scholars have done more in-depth research, but this technology is limited by the number and interval of FBG writing, and cannot achieve long-distance distributed shape measurement. The other direction is the shape sensing based on the distributed optical fiber measurement system. As a medium of shape sensing technology, optical fiber is small in size, light in weight, and has strong electromagnetic interference resistance and corrosion resistance. It can be either a transmission medium or a sensing medium. When the light wave is transmitted in the optical fiber, the optical intensity, phase, frequency and other parameters of the optical fiber will change with the change of environmental parameters such as strain and temperature. The data processing equipment is used to demodulate the modulated light, and then the information of strain and temperature of the optical fiber is obtained. In this paper, the Brillouin scattering in the fiber is used to reconstruct the shape of the fiber or the measured object in contact with it, and the strain change values of more than two fiber cores in the shape sensor are measured at the same time. Then the shape reconstruction algorithm is used to reconstruct the shape of the sensor or the measured object. In this paper, the BOTDA system is built with a spatial resolution of 1 m. A homogenous low-crosstalk seven-core fiber from Changfei Company is selected as the distributed shape sensor. The total length of the fiber is 300 m, the core diameter is 8 μm, the cladding diameter is 150 μm, and the protective layer diameter is 245 μm. The remaining six cores are located at a distance of 42 μm from the middle core and are symmetrically distributed around each other at 60°. At the same time, the seven pigtails of the multi-core fiber are labeled and separated by a fan-in fan-out coupler. By using the BOTDA system, the Brillouin gain spectra of the intermediate core and the off-core are measured, and it is verified that the intermediate core is not affected by bending, and the strain values of each two symmetric off-core are negative to each other. Three unsymmetrical cores with 120° distribution were selected, and the intermediate cores were used as temperature compensation to demodulate the induced variables of each core at different curvature radii. Finally, parallel transmission frame shape reconstruction algorithm is used to reconstruct the shape of seven-core fiber when the curvature diameter is 0.112 m and 0.052 m. When the curvature diameter is 0.112 m, the curvature reconstruction error is 0.375%, which is mainly due to the low spatial resolution of the construction system and the torsion problem in the winding process. Distributed fiber shape sensing technology has a very large application prospect, but there are still many technical difficulties that need to be overcome by researchers. The work in this paper has laid the research foundation for the subsequent distributed fiber shape sensing, and has certain practical significance.