The interaction mode in a mobile layout system was studied. An evaluation experiment was designed and conducted to compare five interaction modes: single pen-based, pen + voice, pen + keyboard, mouse + voice, mouse + keyboard. Subjects were instructed to fulfill the basic tasks of a layout plan de- sign system, such as locating, shaping the property of, and modifying the coor- dinate setting of an object. In plan setting, modification time, accuracy rate and subjective preferences a pen-based system proved the best, pen + voice fol- lowed. This experiment confirmed that different availability of interaction modes resulted in different interaction efficiency in the mobile layout environ- ment. The experiment also provided information on the possible efficient com- bination of interaction modes. This paper gives statistical support to indicate that use of pen is preferable to voice in setting and modification procedures in mobile technology.