Background and Objectives: Systematic reviews (SRs) have been widely applied in traditional Chinese medicine (TCM) and the number has increased markedly since 2003. In order to improve the quality and avoid incorrect conclusions, some topics different from western medicine need to be addressed for preparing SRs of interventions in TCM. Methods: Summarizing the existing problems (reviewing 107 SRs of TCM) and practical experiences, several important factors were identified. Results: Three fundamental points on SRs of TCM were elicited: 1) Treatment classification: there are several types of treatment and different preparations mixed with different herbs in TCM. Chinese herbal medicine can be divided into three categories: Chinese patent medicines, classical prescriptions and clinicians’ self-prescriptions. So, a systematic review titled ‘TCM/Chinese herbal medicine for a certain disease’ is not appropriate as ‘western medicine for a certain disease’. The broad treatment of a review will cause problems in literature search, synthesis analysis, updating and the conclusions will be less clear. 2) Zheng differentiation-treatment: this is at the core of clinical practice of TCM, both for diagnosis and treatment. If authors don’t pay attention to this characteristic in ‘types of patient’ and ‘data analysis’, it may lead to incorrect results. If possible, a subgroup analysis should be carried out based on the types of Zheng. 3) Literature search: as fully as possible, is a principle of literature search for SRs. To collect trials of TCM, it is not enough to search only English databases because many clinical trials of TCM are published in Chinese. Besides selecting mainstream large-scale databases and some specialized databases, Chinese databases (Chinese Biomedical Database (CBM) and China National Knowledge Information (CNKI)) are also required. Conclusion: Besides basing on the criteria of Cochrane Handbook, preparing SRs of intervention in TCM also need to consider the theory and characteristics of TCM.