In order to accurately predict the dimensional error of the machined face in end milling process, prediction model of dimensional error was established by using dynamic milling force convolution model and introducing concept of surface generation window. The influences of deformation errors of workpiece and tool, and spatial error of machine tool and tool runout were considered. It was validated by end milling experiment, that the model was able to correctly predict dimensional error and its distribution range of the workpiece. Moreover, increasing radial depth of the tool will not have a significant impact on the dimensional error of workpiece with the condition of same axial milling depth of cut, spindle speed and feed rate.