In order to achieve precise intrusion detection in fiber optic networks, a fiber optic network intrusion detection method based on Markov decision process is proposed. By using frequency domain partitioning technology to purify fiber optic network signals, the empirical mode decomposition method is used for initial detection of intrusion signals, and the fuzzy analytic hierarchy process is used to determine the credibility of network access behavior. For access behaviors with higher credibility, they are directly passed through, while the remaining access behaviors are judged using Markov decision process, thus achieving intrusion detection. The experimental results show that this method can quickly and accurately detect intrusion signals, especially for the intrusion eavesdropping behavior suffered by the lending dataset, with a detection rate of up to 0.985. In the entire experiment, the minimum detection rate of this method can also reach 0.920, and the maximum average detection misjudgment rate and average detection omission rate are 0.01 and 0.02, respectively. This indicates that the method significantly improves the security and stability of fiber optic networks, providing strong support for ensuring network security.