Data mining is a job in the form of collecting and using data to get a rule, pattern, or relationship in large data. The output of this data mining can be used to facilitate future decision-making. The purpose of this study is to predict the level of sales of computer equipment to make it easier for sellers to meet consumer needs. The data processed in this study include several factors which will later be included in the Roughtset method. The method used is a Rough set. Factors include the name of goods, warranty, price, and level of sales. These factors will later be grouped in the Equivalence Class, where the same attribute values will be grouped. Then proceed to the next stage, namely the Discernibility Matrix which contains a collection of condition attributes that have different condition values. After that, proceed to the Discernibility Matrix Modulo D stage where the columns in the matrix are filled with a collection of condition attributes that have different conditions and different decision values. The next stage is Reduction, which is how to get the condition attributes used to get output in the form of knowledge. The last stage is knowledge which is the result of the reduction obtained. Then the results of the rough set application will be entered into the Rosetta application. The results obtained using the rough set method on 10 samples of computer equipment sales data, obtained 17 new rules or knowledge that can be used as guidelines in decision making to identify the level of computer sales.