This study focused on supply chain planning with multi-echelon, single-product, and multi-period unbalanced system, took into consideration of customer demand, supplier relationship and capacity limit. We proposed a hybrid methodology based on the supplier relationship management (SM), the analytic network process (ANP) and an improved genetic algorithm (GA) to find out optimal supply chain network types and production distribution model. The improved GA called IGA is based on the multi-objective GA and adjustment mechanism. To confirm the usefulness, finally, IGA is used to solve the multi-objective optimization mathematical model of multi-echelon unbalanced supply chain planning in comparison with basic genetic algorithm (BGA).