Calibration and linearization are two important topics that must always be considered to assure the accuracy of measuring systems. Measurement errors, namely offset, gain and linearization errors, can be compensated as long as timely calibration routines are performed in the measurement system. Nowadays, with the advent of smart sensors, the new capabilities associated with microprocessor or microcontroller devices can support new and advanced calibration and self-calibration algorithms that contribute to increase the measurement’s accuracy. In the present paper, an adaptive self-calibration algorithm for smart sensors’ linearization is proposed. The algorithm takes into consideration the probability density function of the measured data in order to reduce the number of calibration points, and of the associated calibration time, for a required level of accuracy. The progressive polynomial interpolation method is considered in order to preserve the values of calibration coefficients, already evaluated for previous calibration points, without starting the algorithmic calculation of a new set of the calibration coefficients for each new additional calibration point. Some simulations and an experimental result, for a square root characteristic of a venturi type airflow transducer, will be presented in order to validate the theoretical expectations.