For optimization of a hydrogen network, a steam reformer is associated to the feedstock and linear programming (LP) is applied. The investigated network consists of one steam reformer and two feedstocks. By exerting LP and the mentioned association, total annual cost decreasing is achieved in a case study in which natural gas and off-gas were considered as feedstocks. The optimization problems of the hydrogen network comprise the hydrogen network retrofit design and the feedstock selection with respect to their cost. Nonlinear programming (NLP) and mixed-integer nonlinear programming (MINLP) models are developed for optimization based on a two-case study: for the first one, an existent optimization method on hydrogen networks is investigated and for the second one, revision of a recent optimization method on hydrogen networks associated by an LP model in the steam reformer unit is applied. These two cases resulted in total annual cost reductions of 34?% and 45.9?%, respectively.