The problem at PT. TMCI is that the lab part is still confused in determining the quality of some cocoa beans, because there is no average calculation. So that the quality of cocoa depends only on the specified parameters, and the moisture content, the count of cocoa beans as the first determinant of quality. The purpose of this study is to make a decision model in a decision making with the Multidimensional Scaling method, and a decision support system as a decision maker in determining the quality of cocoa beans, and implementing a decision support system and calculation of MDS metrics with several parameters, namely based on moisture content, moudly, broken beans, weight of waste, number of cocoa beans and flat cocoa beans. Then apply the MDS method in an Android-based application system, so that from the calculation results can determine which quality is good and which will be mixed again. Data collection is done by observing conditions directly, and conducting interviews. From the results of testing implementation, interviews and testing applications, that the application works properly when used. In terms of appearance and function the application can help the lab and the owner at the time of supplying supplier data. So it was concluded that this application could be well received and answer the problems of the lab and the owner.