Since productivity and product quality are always regarded as important issues in manufacturing technologies, a reliable method for predicting machining errors is essential to meeting these two conflicting requirements. However, the conventional roughness model is not suitable for the evaluation of machining errors for highly efficient machining conditions. Therefore, a different approach is needed for a more accurate calculation of machining errors. This study deals with the geometrical surface roughness in ball-end milling. In this work, a new method, called the ridge method, is proposed for the prediction of the machined surface roughness in the ball-end milling process. In Part I of this two-part paper, a theoretical analysis for the prediction of the characteristic lines of the cut remainder are generated from a surface generation mechanism of a ballend milling process, and three types of ridges are defined. The trochoidal trajectories of cutting edges are considered in the evaluation of the cut remainder. The predicted results are compared with the results of a conventional roughness model.