Abstract: Wideband radar utilized wideband signals to enhance the range resolutions of tracking and generated high-resolution images containing the feature information of targets, from which the targets could be precisely measured, classified and identified. The determination of the detection threshold of wideband radar detector was an important issue in the field of wideband detection. This paper introduced a threshold setting method based on importance sampling, especially for wideband radar detectors that couldn’t or were difficult to determine analytical threshold. First, importance sampling, which could be used to significantly reduce the estimation variance, was introduced to improve the sampling efficiency of rare samples. Then the relationship curve between wideband detector’s false alarm probabilities and thresholds were drawn through accumulating sample weights that passed the thresholds. Finally, the interested thresholds were obtained by linear interpolating the above curve. Simulation experiments showed that the proposed method could effectively enhance the computing scale, and improve the detector threshold computing accuracy and speed.