The retail industry is currently growing rapidly, especially in Indonesia. One form of the retail industry is modern retail which includes supermarkets, minimarkets and others. This study focuses on the grouping of products sold at minimarkets. This research is caused by seeing the phenomenon of the large number of transactions that occur in one day, the result is the number of products sold. This makes it difficult for minimarket managers to determine the next product procurement. Therefore, This study is conducted to group the products sold so that the products that need to be procured are seen next. This study propose a software to perform the grouping using the K-means algorithm. For the data sample, this study obtained sales transaction data for 3 months from the Sastra Mart minimarket. In this study, manual calculations were carried out on 10 samples of beverage data taken randomly from sales transactions which would be divided into 3 clusters. The results of manual calculations, there are 3 drink data entered into the “Sangat Laris” cluster, 2 drink data entered the “Laris” cluster and 5 drink data entered the “Kurang Laris” cluster. The software produced from the research gives the same results as manual calculations in classifying products. This study has also carried out software testing to test all its functionalities, from the test results, everything runs normally and as expected.