Motion control is an important task in industrial automation systems. And the exact motor speed estimation is needed for precise motion control. To obtain the motor speed, linear hall sensor is used in this paper for implementation as a low cost and a simple calculation. Since the linear hall sensor output is sinusoid wave, the measurement equation can be modeled with a sinusoid signal easily. Based on the model, the instrumental variable (IV) method is proposed to estimate the motor frequency in this paper. To prove its performance, the estimation from IV is compared with those from the nominal least squares (NoLS), weighted robust least squares (WRLS), and true value. Experimental results show that the IV method is superior to the NoLS algorithm, and similar with the WRLS algorithm. Moreover, the proposed IV method is useful because it can be applied even if the stochastic properties are unknown or not exact.