In infrared target tracking, the pyroelectric infrared (PIR)sensor can detect the relative direction of a thermal moving target through optical modulation, but cannot get its accurate distance. For this, a swarm intelligence search approach is proposed. At the swing state, PIR can feel the moving target, at the same time, the detection angle is recorded. By analyzing the relationship between the detection angle sequence and the discrete points of moving target trajectory in polar coordinates, finding that the target trace is determined by the polar radius sequence. Multi-dimensional polar diameter particle swarm optimization space is established. Meanwhile, a fitness function to evaluate the particle′s merits is proposed. The target trajectory can be obtained. The proposed method is an effective solution to enlarge the detection range and improve the accuracy of target tracking. Human body′s tracking experiment is performed. The average localization error is not more than 0.4 m. The findings demonstrate the effectiveness of the method.