The increased requirement of hydrogen in the refinery is fulfilled by proper distribution of available hydrogen in the refinery, using additional hydrogen production system and by importing. Amongst these options first one is cost effective as no addition source is required. Thus, the present paper deals with optimum distribution of hydrogen. For this purpose mathematical models are developed based on pressure constraints, source and sinks constraints, compressor flow rate recycle and purity constraints, flow combinations, hydrogen consumption, operating cost, capital cost, payback period etc. The model is developed in stages to show the improvements. Two case studies, Cases A and B, are considered in this paper and for these linear programming (LP), nonlinear programming (NLP), mixed-integer linear programming (MILP) and mixed-integer nonlinear programming (MINLP) models are developed. The results predicted using these models are compared to select the best network. The optimized network for Case A consumes 20.9% less hydrogen, whereas, for Case B it is reduced by 32.3%. (C) 2010 Elsevier Ltd. All rights reserved.