Objective: To explore the medication principle of treating obesity and hyperlipidemia with Chinese herbal medicine by a novel text mining approach. Method: We proposed a data mining algorithm to retrieve simple and meaningful networks from large data sets of obesity and hyperlipidemia. We transfered XML type data sets to the structured database of Microsoft SQL 2 000 and visualized them into different graphs by software Cytoscape version 2.7 in order to detect the potential medication principles. Result: The most common used Chinese herbal medicine applied for treating obesity commonly proved to be Radix Salviae Miltiorrhizae

Fructus Crataegi Pinnatifidae

Rhizoma Alismatis

Radix Astragali Mongolici

Radix et Rhizoma Rhei Palmati

poria

Rhizoma Atractylodis Macrocephalae

Rhizoma Pinelliae;the core couplet medicine is Radix Astragali Mongolici

Rhizoma Atractylodis Macrocephalae

Radix et Rhizoma Rhei Palmati;the most common used herbal medicines applied for treating hyperlipidemia are Radix Salviae Miltiorrhizae

Fructus Crataegi Pinnatifidae

Rhizoma Alismatis

Radix Astragali Mongolici

Radix et Rhizoma Rhei Palmati

Radix Polygoni Multiflori;the core couplet medicine is Radix Salviae Miltiorrhizae and Fructus Crataegi Pinnatifidae. Conclusion: although the primary prescriptions for treating obesity and hyperlipidemia share some common herbs

the cores of the two formulae are significantly distintive

which indicates that the pathogenesis of the two diseases are different according to the theory in Chinese medicine. The novel text mining technology could be a promising approach in the medication principle detecting research of Chinese medicine.