The need for a prediction system of rice production in Barito Kuala Regency, the biggest contributor of rice in South Kalimantan is a reason for to make a research about the prediction of the rice production number. Specify of the number prediction of rice production in large numbers never be done by a manual system because it’s need a long time. Forethere,it’s need an algorithm that can to determine and to predict the number of the rice production automatically. This research method using the K-Nearest Neighbor because it has the principle of data searching with an identic number of K. Search for similarity data of data training which obtained from the Agricultural Department of Barito Kuala Regency with the testing data that included by user. The system with the K-Nearest Neighbor can make avalue precision of 90% and 50% recall. From these research, the system can to predict the number of production well. Key Words: Decision Support Systems, K-Nearest Neighbor, The Rice Production Number, Delphi, Information Retrieval.