Currently, maritime target detection has been widely used in the civil and military fields. However, the difficulty of target detection is increased due to the large fluctuations of seawater, the small imaging area of the target, and the insignificant features. To eliminate these problems, a Gaussian mixed-Markov random field target detection model is proposed, which uses the Gaussian mixed model to adjust the learning rate and to suppress the interference of the dynamic background. Then, using the result of the Gaussian mixed model to calculate the infrared image as the observed value of the Markov random field, a Markov random field model is established. The Markov random field model can solve the existing defects of the Gaussian mixed model. The experimental results show that the method in this paper can achieve good detection results.