In this paper, the problem of detecting a multichannel signal in the presence of spatially and temporally colored disturbance is considered. By modeling the disturbance as a multi-channel auto-regressive (AR) process with a random cross-channel (spatial) covariance matrix, two knowledge-aided parametric adaptive detectors are developed within a Bayesian framework. The first knowledge-aided parametric detector is developed using an ad hoc two-step procedure for the estimation of the signal and disturbance parameters, which leads to a successive spatio-temporal whitening process. The second knowledge-aided parametric detector takes a joint approach for the estimation of the signal and disturbance parameters, which leads to a joint spatio-temporal whitening process. Both knowledge-aided parametric detectors are able to utilize prior knowledge about the spatial correlation through colored-loading that combines the sample covariance matrix with a prior covariance matrix. Computer simulation using various data sets, including the KASPPER dataset, show that the knowledge-aided parametric adaptive detectors yield improved detection performance over existing parametric solutions, especially in the case of limited data.Index Terms-Bayesian inference, generalized likelihood ratio test, knowledge-aided process, multi-channel auto-regressive model, space-time adaptive processing.