Maintaining a low detection probability by defense radars in vital areas is a fundamental requirement for aircraft route planning during aerial surveys. This necessitates considering and assessing the variations in radar detection probability caused by the kinematic state uncertainty of the aircraft itself. To this end， analysis method based on unscented transformation （UT） is proposed to evaluate the impact of aircraft kinematic state uncertainty on detection probability. Relevant models regarding aircraft detection by radar are presented. The UT method is introduced to establish an process for the propagation of aircraft kinematic state （involving position， and attitude） uncertainty and their impact on radar detection probability. A simulation experiment scenario is designed wherein an aerial survey aircraft is detected by a single-pulse defense radar. The Monte Carlo method is employed to verify the high analysis accuracy of the UT method. Additionally， the patterns regarding how varying degrees of aircraft kinematic state uncertainty affect detection probability are analyzed. Furthermore， comparisons with the traditional linear covariance analysis method demonstrate that the UT method possesses both simplicity and high computational efficiency.