A new thin film design and optimization method based on clustering and global optimization algorithms is presented,which can iteratively change the initial structure.The design solution is calculated more thoroughly by this new design method than the traditional local optimization technique.When initial film structure is far from the design target,this kind of clustering global optimization method has the unique advantage that a satisfied consequence can still be obtained.A practical thin film design task which is a broadband antireflection coating from the visible light to short- wave infrared is used to testify the ability and superiority of the clustering global optimization method.