In this paper, an effective neural acquisition system is suggested for acquisition of the DS/SS communication system. The suggested system uses a binary neural network which geometrically analyzes and learns a given PN sequence in the binary field. the probabilities of detection and false alarm are obtained and are compared to simulation values. The equation of the mean acquisition time is derived and is compared to the doubledwell time of the serial serial search system. The significant improvement of the performance is demonstrated. As the length of synchronization sequence becomes longer and the number is increased, the performance is improved.