Due to the complex control structure and lack of support from the main grid, small signal stability has always been one of the key issues in studying the operation of microgrids in islanded mode. In order to investigate the sensitivity of frequency oscillations to system parameters and quantify the influencing factors, the eigenvalue sensitivity algorithm is employed in this study. A multi-SST (Solid-State Transformer) ring microgrid with virtual impedance is designed to establish a complete small signal model for the microgrid. Different oscillation modes are determined based on the imaginary parts of the eigenvalues, and the key parameters of different oscillation modes are studied through eigenvalue sensitivity analysis. The research results show that the influence trends and magnitudes of system parameters on oscillation modes are different. In high-frequency oscillations, the SST voltage loop control parameters and SST output voltage have the largest impact, followed by SST filter current. In sub-synchronous oscillations, the SST output current has the greatest influence, followed by network line parameters. In low-frequency oscillations, the active-power frequency droop coefficient of the SST power control loop has the largest impact.