scButterfly is a versatile single-cell cross-modality translation method based on dual-aligned variational autoencoders and data augmentation schemes. It addresses the challenges of noisy multi-modal data and limited sample sizes by preserving cellular heterogeneity and revealing cell type-specific biological insights. The method is superior to baseline methods in various settings, including diverse sequencing protocols, unpaired data, and batch effects. scButterfly enables integrative multi-omics analysis, data enhancement, and automatic cell type annotation, demonstrating its broad applicability and potential for predicting single-cell perturbation responses.scButterfly is a versatile single-cell cross-modality translation method based on dual-aligned variational autoencoders and data augmentation schemes. It addresses the challenges of noisy multi-modal data and limited sample sizes by preserving cellular heterogeneity and revealing cell type-specific biological insights. The method is superior to baseline methods in various settings, including diverse sequencing protocols, unpaired data, and batch effects. scButterfly enables integrative multi-omics analysis, data enhancement, and automatic cell type annotation, demonstrating its broad applicability and potential for predicting single-cell perturbation responses.