Food is a basic need that must be fulfilled and easily accessible to the entire community. After the end of the pandemic period, it still caused several sectors to decline, including the agricultural sector, which resulted in crop yields also declining. The problem faced by several regions in Indonesia, one of which is the North Sumatra region, is that the availability of food products has decreased and increased unstably due to the lack of information about the grouping food security every year. This results in the food needs of the people in each region being unfulfilled. The purpose of this study is to group areas with the number of increases and decreases in food crop yields in North Sumatra using the K-Medoids algorithm. The K-Medoids algorithm includes a deflection algorithm that is quite efficient in carrying out the shaking of small datasets and the search for the most representative points and can overcome outliers. So that it can be used in the floundering of the influence of productivity and the level of food security. The results showed that the application of the K-Medoids algorithm resulted in a DBI (Davies Bouldin Index) value of 0.062 and a Silhouette Coefficient value of 0.8980, with the number of clusters as many as 3 clusters where Cluster_0 dominated by corn food crops experienced an increase in production by 5% and peanuts by 5%, Cluster _1 was dominated by a decrease in the number of soybean production yields by 38%, and Cluster_2 dominated by a decrease in green bean yield by 33%.