Indonesia is a tropical country that has a diversity of plant species. Plants consist of three basic organs namely roots, stems and leaves. The leaves are one of the parts that are often used to classify plants, since each type of plant has different characteristics. The shape of leaf edges can be used for reference in leaf classification. The human brain has limitations in processing or extracting information on the types of plants based on leaves. Therefore, it requires the transfer of manual knowledge to a digital system. So in this study made a system that is able to do the introduction of leaves using the extraction of features on the leaves using linear metide Discriminant Analysis ( LDA) and minkowski distance. The process of controlling leaf image patterns begins with the taking of leaf imagery, then goes to the pre-management stage to distinguish objects from the background. After that it goes into the feature extraction stage using Linear Discriminant Analysis (LDA) to obtain characteristic features of the imagery and Minkowski Distance to perform an introduction of the leaf pattern.Based on the results of the study with the amount of data as many as 40 classes with each class as many as 6 images, with a trained image of 160 leaf imagery and test imagery as many as 80 leaf imagery. At the time of introduction using distance Minkowski used 3 coefficients namely minkowski coefficients 1, 2, and 3. From ¬¬each coefficient minkowski obtained a percentage of accuracy. Minkowski 1 coefficient of 41.25%, minkowski 2 coefficient of 33.75%, and minkowski 3 coefficient of 30%. The percentage of accuracy in this study can not produce an estimated 80% because the amount of data greatly affects the percentage result, the more data there is then the percentage value will also be smaller.