Conventional Rayleigh wave suppression methods take advantage of differences in frequencies or apparent velocities between surface waves and effective waves to separate Rayleigh waves. However, the suppression may fail when the waves present small differences. To solve this problem, the super-virtual interferometry method is introduced to predict and suppress Rayleigh waves. It is a data-driven method, which is independent of frequencies or apparent velocities, and presents good applicability to complex near surface. Compared with conventional interferometry method which only predicts inter-receiver Rayleigh waves and requires a real shot around the receiver position, the super-virtual interferometry method is able to achieve shot-receiver Rayleigh waves, followed by adaptive subtraction using multichannel L₁-norm. The synthetic data examples demonstrate the validity of the proposed method and obtain good results.