A novel wideband direction-of-arrival estimation algorithm based on improved generalized regression neural network(IGRNN)and principal component analysis(PCA) is proposed in order to obtain high estimation precision.Firstly,the PCA method is used to lower the dimension of train samples in order to reduce the complexity of GRNN whose parameters are optimized by particle swarm optimization algorithm.At the same time,accurate and rough estimation models are established respectively according to different focusing angles.The accurate model is used for DOA estimation after rough estimation to avoide the influence of focusing angles.Simulation results show that our algorithm has high estimation precision and efficiency.