Since hospital information systems have been introduced in large hospitals, a large amount of data, including laboratory examinations, have been stored as temporal databases. The characteristics of these temporal databases are: (1) in-homogeneity of each record, (2) a large number of attributes in each record and (3) bias of the number of measurements for patients suffering from severe chrononic diseases. The characteristics of these temporal databases are: Even medical experts cannot deal with these large databases, the interest in mining some useful information from the data are growing. In this paper, we introduce a combination of extended moving average method, multiscale matching and rule induction method to discover new knowledge in medical temporal databases. This method was applied to two medical datasets, the results of which show that interesting knowledge is discovered from each database.