Synthesis of heat integrated multi-component distillation systems is complex due to its huge search space for structural combination and optimization computation, To provide a systematic approach and tools for synthesis design and retrofits:of distillation systems, an improved genetic algorithm (IGA) is proposed. The algorithm inherits main ideas in evolutionary computing, employs a distributed sub-population strategy to avoid local optima, and applies for a continuous variable space coding procedure, which is computational fast and table in converging to global optima. This paper describes IGA-based optimization strategy for synthesis and optimization of separation sequences and their heat exchanger network. Two example are presented.