Abstract: The conventional detection algorithms have poor signal detection performance in non-Gaussian clutter, and the signal detection in non-Gaussian clutter is an important issue in radar signal processing field. In this paper, we consider a signal detection method in non-Gaussian clutter using some prior knowledge about clutter. The statistical characteristics of clutter can be modeled as compound Gaussian process, which is the product of speckle components, which can be modeled as complex Gaussian process, and texture components. Based on the Bayesian methods, using the inverse gamma distribution as the prior distribution of texture components, a knowledge-aided signal detection algorithm is proposed. The computer simulations results show that, the knowledge-aided signal detection algorithm outperform the conventional detection algorithms. Furthermore, in this paper, we use IPIX radar sea clutter data from McMaster University as the object of study, and the prior information of non-Gaussian clutter is obtained using maximum likelihood estimation of clutter data. The detection performance of the algorithm is analyzed, and the results show that, the detection performance of this algorithm is outperform the other conventional signal detection methods, and can achieve a better detection performance in sea clutter data with different resolution.