PointPillar target detection algorithm is a mainstream 3D point cloud target detection algorithm, and its fast operation speed is widely used in industry. To address the problem of low detection accuracy of PointPillar target detection network, a PointPillar target detection algorithm based on the attention mechanism is proposed. The algorithm introduces CBAM attention mechanism on the basis of PointPillar to achieve the amplification of local information in the three scale feature maps and better extract the more important feature information. The experimental results show that the PointPillar+CBAM target detection algorithm achieves good detection results, taking the medium difficulty case, the mAP of AOS mode reaches 69.26, the mAP of 3D mode reaches 61.33, and the mAP of BEV mode reaches 68.90.