To correctly split an order among parallel suppliers is one of the most important ways to improve the agility and competitiveness of a supply chain. For this order splitting problem, an evaluation criterion of production load equilibrium among parallel suppliers is introduced and a two objective order splitting model is developed to minimise the comprehensive cost and balance the production loads among the selected suppliers. The procedure of the suppliers' selection is proposed and non-dominated sorting genetic algorithm (NSGA-II) is applied to find the feasible solution set of Pareto. The simulation results indicate that the proposed model and algorithm is capable of obtaining satisfactory solutions.