Objective: This paper uses data mining technology to visually discover the variety specification and common efficacy of Tujia medicine with semi-structured data. Method: Collecting and sorting out the representative medical documents of Tujia such as Chinese National Medicine Dictionary and Tujia Medical Formulary is to establish the original database of Tujia medicine, including Latin scientific name, medicinal position, sexual taste, toxicity, efficacy, and indication of diseases, etc. With the help of Visual FoxPro 9.0, researching the frequency statistics and variety specifications of Tujia medicine are made. With the help of TCMISS (V2.5 ) ,Cytoscape3.6.1, Gephi0.9.2 and other software , researching the semantic relationship between Tujia medicine " medicine - efficacy - indication" is explored and visually displayed. Result: A total of 1453 Tujia medicines with 771 prescriptions were screened, of which Tujia medicines are mainly cold ( cool ) and mild, bitter medicines have the highest frequency, followed by pungent medicines and sweet medicines. The efficacy of clearing away heat and toxic materials, expelling wind and dehumidifying has the highest frequency. Skin sores and tinea, gynecological diseases and so on are the main indications with a greater degree of correlation. Conclusion: The visualization of the hidden relationship between Tujia medicine variety specification and " Property - Taste - Efficacy - Indication" can provide an effective reference. It laid a foundation for the improvement of Tujia medicine information research.