Introduction: Recently identified human homologues of ALKB protein have shown the activity of DNA damaging drugs, used for cancer therapy. Bioinformatics study of hABH2 and hABH3 had led to the discovery of a novel DNA repair mechanism. Very little is known about structure and function of hABH4, one of the members of this superfamily. Therefore, in present study we are intended to predict its structure and function through various bioinformatics tools. Materials and Methods: Modeling was done with modeler 9v7 to predict the 3D structure of the hABH4 protein. This model was validated with the program Procheck using Ramachandran plot statistics and was submitted to PMDB with ID PM0076284. The 3d2GO server was used to predict the functions. Residues at protein ligand and protein RNA binding sites were predicted with 3dLigandSite and KYG programs respectively. Results and Discussion: 3-D model of hABH4, ALKBH4.B99990003.pdb was predicted and evaluated. Validation result showed that 96.4 % residues lies in favored and additional allowed region of Ramachandran plot. Ligand binding residues prediction showed four Ligand clusters, having 24 ligands in cluster 1. Importantly, conserved pattern of Glu196-X-Pro198- Xn-His254 in the functional domain was detected. DNA and RNA binding sites were also predicted in the model. Conclusion and Prospects: The predicted and validated model of human homologue hABH4 resulted from this study may unveil the mechanism of DNA damage repair in human and accelerate the research on designing of appropriate inhibitors aiding in chemotherapy and cancer related diseases.