Most of previous researches in a load sharing system have focused on expressing reliability of a given system under the assumption that the number of components used in the system is pre-fixed. Only one recent research has studied the optimal number of components for maximizing system reliability under one-component-out capability, assuming that the component follows an exponential lifetime distribution and a power function explains the load-life relationship. This paper intends to generalize such a result to consider the general distribution for the component lifetime and the general functional for the load-life relationship. We derive the system reliability in terms of the number of components used in the system. A numerical example is given to illustrate the optimal number of engines to maximize reliability of the liquid rocket propulsion system. The result indicates that a large number of components each withstanding a small load is preferred to a small number of components each withstanding a large load if the component failure rate increases with time and the failure rate decreases fast with the reduction in the design load.