To address the challenge that existing thickness measurement methods struggle to achieve high-precision detection for the overall thickness of discontinuous carbon fiber reinforced polymer (CFRP) circular cell honeycomb core layers, a thickness measurement method based on dual-surface scanning with a line laser sensor and common reference reconstruction is proposed. This method obtains 3D profile point clouds of the upper and lower surfaces of the CFRP circular cell honeycomb core layer through separate scans and introduces a preset common reference to achieve unified thickness reconstruction of the workpiece. To resolve the key difficulty of unstable common reference acquisition caused by the lack of high-precision geometric registration features in the surface point clouds, an internal expansion calibration block is designed to be fixed inside the composite tube. This ensures that the common reference can be scanned stably and reliably before and after the workpiece pose is flipped. Regarding data processing, considering the discontinuous nature, high noise level, and significant local density variations of the CFRP circular cell honeycomb core layer point cloud, a staged denoising strategy based on statistical characteristics and clustering features is constructed. Statistical filtering is used to remove outliers, combined with K-means clustering for adaptive screening of redundant data, thereby improving the stability and accuracy of surface profile reconstruction. To verify the measurement accuracy of the proposed method, comparative tests are conducted on small-scale CFRP circular cell honeycomb core layers using a coordinate measuring machine, with the coordinate measuring machine results serving as the ground truth. The experimental results show that the thickness measurement accuracy reaches 0.03 mm, with a relative error controlled within 0.1%, meeting the precision requirements for thickness detection of CFRP circular cell honeycomb core layers. Furthermore, on-machine measurements of large-scale CFRP circular cell honeycomb core layers are performed, validating the feasibility and effectiveness of the proposed thickness detection method in practical engineering applications.