Abstract: Addressing the issue of degraded detection performance in frequency diverse array-multiple-input multiple-output (FDA-MIMO) radar when detecting constant-speed targets due to signal mismatch, a method to enhance the robustness of the detector is proposed. First, a random variable is introduced into the signal assumption, which is assumed to follow a Gaussian distribution and share the same covariance structure as the noise, while its magnitude is regulated by an unknown robustness factor. Then, based on the one-step generalized likelihood ratio test, two-step generalized likelihood ratio test, Wald and Durbin criterion, four adaptive detectors are devised. Finally, simulation experiments demonstrate that constant false alarm rate properties are satisfied by all detectors. In terms of performance, superior performances are achieved by the one-step generalized likelihood ratio test and Wald detectors under the matched condition, whereas different degrees of robustness are exhibited by the detectors in the presence of signal mismatch. Important theoretical support and methodological options are provided for robust detection of FDA-MIMO radar in practical complex environments by the study in this paper.