Radar is a major piece of object detection equipment and its maximum detection range of radar is a key performance indicator. This indicator is usually assessed by carrying out multiple flight check trials and analyzing the difference between the measured distance and the real distance of the target (known as the one-off error in ranging). However, the existing methods ignore the uncertainties in experimental data. To address this problem, the accuracy of the indicator calculation is improved by data modeling and uncertainty analysis. The entire flight check segment is firstly divided into distance intervals, and then a distribution is fitted to the measured one-off errors for each distance interval. Subsequently, a consistency test is performed on the range of one-off error data at the same distance interval in multiple tests. According to the test results, the type of uncertainty is identified, and the maximum detection range evaluation model of the radar is further established based on the identified type of uncertainty. For multiple trials that pass the consistency test, the maximum detection distance is obtained by extrapolating the detection probability of radar at different distances taking into account only the aleatoric uncertainty. Conversely, both aleatoric and epistemic uncertainties are considered and interval theory is used to extrapolate the predicted detection probability intervals separately, based on which the maximum detection distance is determined. The assessment model is then applied in conjunction with the detection range testing data of a specific radar, and the analysis results verify the reasonableness of the assessment model.