A prevalent mode of optical tweezers involves position clamping that regulates a constant position of a trapped bead. Traditional schemes employ the measured bead position in the open loop or the control signal in the position-clamp mode as an estimate of external force on the trapped bead. This article shows that traditional methods introduce fundamental limitations on bandwidth of the external force estimation. A method is presented that leads to an order of magnitude increase in the bandwidth of the external force estimation. Furthermore, a comprehensive modeling paradigm is introduced that facilitates estimation of forces on the bead.