Multistatic radar has the advantage of spatial diversity, but its detection performance is decreased in dense multipath scattering environments. This problem can be solved by using the time reversal (TR) technique to take advantage of the multipath effect to improve target detection; however, this usually requires a stationary channel response that is difficult to achieve in practice. This article studies a TR detection algorithm for a multistatic radar system in a varying multipath channel environment. We establish a varying channel response model and derive a time reversal likelihood ratio test (TR–LRT) detector to utilise the characteristics of multiple paths when the multipath environment is changing during the detection process. We use Monte Carlo simulations and theoretical analysis to show the superior performance of the proposed TR detector compared with a conventional detector. Our proposed TR–LRT detector shows good robustness to environmental variations. Simulation results show that better detection results are achieved in a more severe multipath scattering environment.