The marine current turbine (MCT) is becoming more and more popular to produce eco-friendly electricity. However, its performance is negatively affected by MCT imbalance fault. In this paper, an integration fault detection method using stator voltage for MCT is proposed. This method uses an integration way to detect the imbalance fault. The proposed method comprises three steps: First, the data conversion is based on Hilbert transform and the extreme value searching, and then the imbalance fault signature extraction based on the frequency sequences subtraction (FSS). Last, to reduce the data dimension and to set the fault detection limit, a data vector selection method based on principal components analysis (PCA) (called preliminary-selection-based PCA (PS-PCA)) is proposed, the adaptive fault detection is realized by calculating Hotelling T2 and SPE (squared prediction error). Finally, a marine current prototype experimental platform was built to verify the proposed methods. The experimental results show that the method in this paper has high detection accuracy in the fault detection of MCT imbalance under the variable flow rate.