Our direct interactions with the environment are performed through the sense of haptics. Touch and kinesthesia are used to extract objects properties’ as well as to control our motion relative to them. The effectiveness of this sensorimotor control is a key question in the field of Human-Computer Interaction to enhance user performance. This is notably the case when we use a touchpad to control a visual cursor on a separate screen (e.g., a laptop). The control of these graphical interfaces is configured through a mapping between the motor space (e.g., the touchpad) and the visual space (e.g., the screen) called Transfer Function (TF). When we use the touchpad to control the cursor on the screen, the motion is compounded of a preprogrammed phase performed at high speed, and a following homing phase, performed at low speed and based on visuohaptic feedbacks (Elliott et al. 2010). Some operating system TFs (e.g., Windows, OS X) are based on this principle with a visuomotor gain during the preprogrammed phase which is high while it is low during the homing phase to reduce the Movement Time (MT). Such TFs have been shown to enhance performance with this increasing visuomotor gain (Casiez et al. 2008; Casiez and Roussel 2011). However, the reasons of this improvement are not totally elucidated, notably because the prescribed gains are non-linear. Here we analyzed the kinematics of a pointing task with a linear velocity- based TFs to assess how we plan and control our movement based on vision and haptics (i.e., touch and kinesthesia involved in motion perception). We compared two non-linear increasing and decreasing TF with constant gain TFs.