A new method based on support vector regression (SVR) is proposed to extract the conductivity of the ferromagnetic thin films in this paper. First, the conductivity function with the thickness of the thin films and the reflection coefficients of the network is constructed by SVR based on 3D electromagnetic (EM) simulation data. Then, the reflection coefficients of the network under test are measured by vector network analyzer. At last, the conductivities of the ferromagnetic thin film are extracted by using the established SVR model from the measured reflection coefficients. The experimental results show that, using the method proposed in this paper, the conductivity of the thin films can be extracted accurately from 100 MHz to 10 GHz.