Fine-mapping causal variants is challenging due to linkage disequilibrium and the lack of interpretation of noncoding mutations. Existing fine-mapping methods do not scale well on inferring multiple causal variants per locus and causal variants across multiple related diseases using a large number of annotations. Moreover, many complex traits are not only genetically related but also potentially share causal mechanisms. We develop an integrative Bayesian fine-mapping model named RiVIERA-MT. The key features of RiVIERA-MT include: 1) the ability to model epigenomic covariance of multiple related traits; 2) efficient posterior inference of causal configuration; 3) efficient Bayesian inference of enrichment parameters, allowing incorporation of large number of functional annotations; 4) simultaneously modeling the underlying heritability parameters. We conducted a comprehensive simulation studies using 1000 Genome and ENCODE/Roadmap epigenomic data to demonstrate that RiVIERA-MT compares quite favorably with existing methods. In particular, the efficient inference of multiple causal variants per locus led to significantly improved estimation of causal posterior and functional enrichments compared to the state-of-the-art fine-mapping methods. Furthermore, joint modeling multiple traits confers further improvement over the single-trait mode of the same model, which is attributable to the more accurate estimation of the functional priors. We applied RiVIERA-MT to 27 GWAS summary statistics, in-cis eQTL of risk genes in 44 tissues from GTEx samples, and finally to infer causal GWAS/eQTL SNPs, causal genes, and causal pathways. To leverage potential tissue-specific epigenomic co-enrichments among these traits, we harnessed hundreds of functional annotations compiled from ENCODE/Roadmap consortium. Overall, we observed an improved power in detecting functionally coherent risk SNPs with the Bayesian prior dictated by the weighted tissue-specific annotations compared to solely genetically driven model.