The oscillation data of a DFIG-based wind farm via the VSC-HVDC grid-connected system contains a lot of information. These can illustrate the comprehensive effect of uncertain factors on Sub-Synchronous Oscillation (SSO). This paper proposes a data-driven method to analyze the influence of the combination of uncertain factors of wind speed and current fluctuation at the Point of Common Coupling (PCC) of the wind farm on SSO. First, the system oscillation data is processed in segments according to wind speed. FFT based on Nuttall window interpolation is applied to detect the SSO component in the power data corresponding to the results of small-signal analysis, and the amplitude is extracted. Then, a Gaussian Mixture Model (GMM) clustering algorithm is adopted to cluster the factor combination, and the clustering quality is evaluated by three internal validity indicators. Finally, the influence of wind speed/current fluctuation clustering on SSO is analyzed from the perspective of the amplitude change of the SSO component. The results show that in the aspect of factor combination clustering, the proposed GMM clustering method has a better clustering effect than K-Means. When the combination of wind speed/current fluctuation factors belongs to partial clustering, the SSO of the system will worsen. This work is supported by the National Natural Science Foundation of China (No. 51977128).