The research carried out aimed to produce two model / equation, which are the surface roughness and the machining vibration. The generated machining vibration model can be used as substitution parameter to predict surface roughness.Workpiece material used was aluminum-alloy. Aluminum alloys was processed on CNC machines of TU-3A. The first phase of the study (first year), the focus on the synthesis and study of theoretical models to determine the optimum level of accuracy that can be achieved from the model.Design and research methods are divided into several stages. Starting from preparation, cutting parameter settings, machining process, measurement, until to processing and data analysis. Based on the cutting parameters that have been designed before it, then the machining process is performed to obtain the data. Data obtained from the measurement of machinery vibration and surface roughness. Overall, the amount of data obtained from the machining process as many as 125 rows of data. Data processing done to obtain surface roughness and machining vibration equations coupled with the accuracy of each model. Model produced in this study were 5 (five) model, both of surface roughness and machining vibration. But from 5 (five) models, only selected 1(one) model which has an optimum level of accuracy. In this study produced each 1 (one) model of surface roughness and 1(one) model of machining vibration couple with the accuracy of each model.The resulting model for surface roughness has a level of accuracy (theoretically) of 83,533%. While the model generated for machining vibration has the highest accuracy (theoretically) of 82,188%. Keywords: Spindle speed, feedrate, depth of cut, vibration machining, and surface roughness.