Two experiments were conducted to compare target pointing performance with a pen (stylus) and with a cursor key. on small displays. In experiment 1, we examined participants’ performance of target pointing with both input methods at different target sizes. It was found that pen operation is more erroneous than key based operation whn target size is smaller than 5 mm, but at a target size of 5 mm, the error rate decreased to the same level as for key input. In experiment 2, we examined the effect of the number of targets. The results showed, with a target size of 5 mm, the pen could point to targets quicker than with key input, when the distance to the target exeeds a path length of 3 steps.