The number of biomedical literatures is growing rapidly; most of these literatures are available online in an unstructured textual format, such as in the PubMed online system which is a tremendously rich information source. Reading every available article is a time-consuming task in addition to the fact that researchers have to face information overload problem, and that the coverage of databases that provide information manually extracted from the literature is limited. In this paper, we propose a text mining based system, which is Bio Text Mining System (BTMS), which focuses on extracting information automatically from computer-readable literatures and discovering knowledge from these publications in order to identify biological terms in these publications. The BTMS is a rule-based system, which based on defining a set of rules to extract biological terms in different styles of research publications. For example, a rule may read like “tag all tokens that consist of capital characters as named entities”. Rules are usually generated by humans, since such systems use rules to identify gene names, they solve the problems of new names if these new names contain some characteristics as the same as those names that considered in rules. The BTMS easily adds new rules to increase system performance, also easily discovers what rules tag the wrong names after observing the tagged text. As a result, this would improve the performance of system to support biomedical researchers to discover more and more knowledge in their domain. The proposed system is capable of mining the PubMed online system in an interactive manner. It achieves an F-score 90.286% on the Yapex corpus.