Community structure is an important property to uncover structural and functional features in various complex systems.In this paper,we propose an improved spread algorithm based on Principal Component Analysis (PCA) to detect overlapping community structure in the complex network.The proposed algorithm uses PCA to choose the optimal number of eigenvectors self-adaptively,then calculates the Laplace matrix and maps nodes into low dimension subspace.At last,the FCM algorithm is used to reveal the overlapping community structure.The simulation results in real world and artificial networks show that the proposed algorithm can detect reasonable overlapping communities in the complex network.