The condition of an offshore wind turbine (OWT) should be monitored to assure its reliability against various environmental loads and affections. The modal parameters of the OWT can be used as an indicator of its condition. This paper combines the Kalman filter, the random decrement technique (RDT), and the stochastic subspace identification (SSI) methods and proposes an RDT-SSI method to estimate the operational frequency of an OWT subjected to ambient excitation. This method imposes no requirement on the input/loads; therefore, it is relatively easy for field application. An experimental study with a small-scale OWT was conducted to verify the accuracy of the proposed RDT-SSI method. The test results implied that the frequency estimated by the RDT-SSI method is close to that estimated by an impact hammer test. Moreover, the small-scale OWT was buried at different embedment depths to simulate the influence of the scouring phenomenon, and the frequency of the OWT decreased with decreasing embedment depth. Additionally, the bolts at the root of the turbine blades were also loosened to investigate their influence on the frequency. As more blades were loosened, the identified frequency of the OWT also decreased, indicating that the proposed RDT-SSI method can be employed for the health monitoring of an OWT.