For mining and inheriting the traditional Chinese medicine (TCM) experience, it is of great importance that information technology is used to analyze the TCM prescriptions automatically. In this paper, the fuzzy clustering method based on perturbation theory was adopted to analyze the TCM prescriptions. First the character, taste, meridian affinity and toxicity of each single medicine in a specific TCM Prescription were processed with fuzzy quantization. Then the fuzzy clustering based on transitive closure matrix of perturbation theory was applied to analyze the prescription composing principles. As an instance, the minor bupleurum decoction was specifically analyzed by such clustering method, and the comparison was made between pure transitive closure method and transitive closure method based on perturbation theory.  Experimental results indicated that clustering based on the perturbation theory was capable of revealing the composing principles of TCM prescription more objectively and outperforms the fuzzy clustering based on traditional transitive closure matrix, which had sound guiding significance for prescription analysis.