Muscle spindles provide critical information about movement position and velocity. They have been shown to act as stretch receptors in passive muscle, however, during active movements their behavior is less clear. In particular, spindle responses have been shown to be out-of-phase or phase advanced with respect to their expected muscle length-sensitivity. Whether this apparent discrepancy of spindle responses between passive and active movements is due to fusimotor (γ-drive) remains unresolved, since the activity of fusimotor neurons during voluntary non-locomotor movements are largely unknown. We developed a computational model to predict fusimotor activity and to investigate whether fusimotor activity could explain the empirically observed phase advance of spindle responses. The model links a biomechanical wrist model to length- and γ-drive-dependent transfer functions of type Ia and type II muscle spindle activity. Our simulations of two wrist-movement tasks suggest that (i) experimentally observed type Ia and type II activity profiles can to a large part be explained by appropriate, i.e. strongly modulated and task-dependent, γ-drive. That (ii) the empirically observed phase advance of type Ia or of type II profiles during active movement can be similarly explained by appropriate γ-drive. In summary, the simulation predicts that a highly task-modulated activation of the γ-system is instrumental in producing a large part of the empirically observed muscle spindle activity for voluntary wrist movements.