Sensor/system calibration is a crucial step in every measurement system. To perform the calibration the system/sensor input is changed, and its output is measured. The device calibration is obtained from the measured input and output values. When the input/output relation is a straight line, a linear regression is the most common method to obtain the relation. However, this method does not consider uncertainties in the system/sensor input. In this paper, an alternative process based on the Monte Carlo method is proposed. This method can be used considering the uncertainties of the measured inputs and outputs and is flexible to deal with situations for any type of probability density functions of the measured values.