The development of technology that is advancing rapidly today encourages the emergence of buying and selling processes that can be done online. The many conveniences that are felt in the process have made many people switch from buying and selling conventionally to buying and selling online. Clothing is one of the primary human needs that do not escape online sales. However, the obstacle experienced was when choosing the size, because when buying online the buyer could not try on the clothes, thus creating doubts in choosing the size of the clothes that matched the buyer's body size. Therefore, this study develops a smartphone-based application that is used to recommend clothing sizes. The stage that is passed to get the results, namely, the data training process will be carried out on the dataset used. Furthermore, it takes determinant variables that affect the size of a person's clothes, namely gender, weight, height, and body shape to be able to make predictions using previously trained datasets. The result of this study is a smartphone-based application that is useful for recommending clothing sizes. The test results using the Confusion Matrix for 21 test data taken randomly from 207 training data, showed an accuracy rate of 67%.