In the process of radar signal detection, the entire-section pulse compression detection of the long-period signal will be time-consuming and take up too many resources. A segment matching algorithm is used to achieve local pulse compression detection. First, the received echo signal is segmented according to certain rules, and the sliding window detection is carried out in each segment after filtering the data segment by segment, and then the result of the first judgment is saved after data screening. After a number of pulse cycles, the preserved data is aligned in rows. The second judgment is obtained after fine processing, and the target results are ultimately determined. Simulation and test results show that, compared with the traditional method of first compression and then accumulation detection, this method can effectively filter out lots of useless information in processing and save resources, and it is easy to implement.