Chromatin features are widely used for genome-scale mapping of enhancers. However, discriminating active enhancers from other cis-regulatory elements, predicting enhancer strength, and identifying their target genes remains challenging. Here we establish histone H2B N-terminus multisite lysine acetylation (H2BNTac) as a genuine signature of active enhancers. H2BNTac prominently marks candidate active enhancers and their target promoters and discriminates them from ubiquitously active promoters. Two mechanisms afford the distinct H2BNTac specificity. (1) Unlike H3K27ac, H2BNTac is specifically catalyzed by CBP/p300. (2) H2A-H2B, but not H3-H4, are rapidly exchanged through transcription-induced nucleosome remodeling. H2BNTac-positive candidate enhancers show a high validation rate in orthogonal enhancer activity assays, and a vast majority of endogenously active enhancers are marked by H2BNTac and H3K27ac. Notably, H2BNTac intensity predicts enhancer strength and outperforms the current state-of-the-art models in predicting enhancer target genes. These findings have broad implications for generating fine-grained enhancer maps and modeling enhancer-dependent gene regulation.