An integrated strategy based on experimental designs for the development, optimization and validation of the fingerprint method of Sheng-Mai-San by MEKC has been described. Orthogonal and sequential uniform designs were employed to select important experimental parameters and optimize CE conditions. Method validation was performed in terms of injection precision, sample stability test and robustness testing. Additionally, conventional modeling method was used to predict the optimum separation conditions for comparative purpose. The strategy described can also be utilized for fingerprint development in the quality control of other herbal medicines.