We investigated whether corticospinal excitability during motor imagery of actions (the power or the pincer grip) with objects was influenced by actually touching objects (tactile input) and by the congruency of posture with the imagined action (proprioceptive input). Corticospinal excitability was assessed by monitoring motor evoked potentials (MEPs) in the first dorsal interosseous following transcranial magnetic stimulation over the motor cortex. MEPs were recorded during imagery of the power grip of a larger-sized ball (7 cm) or the pincer grip of a smaller-sized ball (3 cm) with or without passively holding the larger-sized ball with the holding posture or the smaller-sized ball with the pinching posture. During imagery of the power grip, MEPs amplitude was increased only while the actual posture was the same as the imagined action (the holding posture). On the other hand, during imagery of the pincer grip while touching the ball, MEPs amplitude was enhanced in both postures. To examine the pure effect of touching (tactile input), we recorded MEPs during imagery of the power and pincer grip while touching various areas of an open palm with a flat foam pad. The MEPs amplitude was not affected by the palmer touching. These findings suggest that corticospinal excitability during imagery with an object is modulated by actually touching an object through the combination of tactile and proprioceptive inputs.