Abstract: In astronomical light curve period extraction, the presence of strong noise often requires a traditional algorithm to perform extensive signal preprocessing, which reduces detection efficiency. To improve period detection efficiency, an innovative weighted limited penetrable visibility graph model is proposed based on complex network theory. By transforming the astronomical light curve signals into complex networks, the proposed method enhances noise immunity while preserving intrinsic signal characteristics. After that, the weighted limited penetrable visibility graph model is obtained by applying Mahalanobis distance weighting to the adjacency matrix of the limited penetrable visibility graph. Finally, the accurate light curve period is extracted using the method of the graph Fourier transform. Experimental results on real data show that this method can accurately extract periodic components while improving detection efficiency.