A new cooperative spectrum sensing (CSS) method based on low-rank symmetric subspace clustering is proposed to improve spectrum sensing performance in complex environments. To align with the evolution of cellular networks and leverage spatial observations, a CSS model with multiple primary users and antennas is considered. To address the asymmetric weighting issue of data pairs, a low-rank representation with a symmetric constraint is proposed that reduces the signal dimensions. Finally, features are extracted from the symmetric coefficient matrix and clustered using the K-means algorithm. Through simulation, the effectiveness of the novel algorithm is numerically validated through comparative experiments.