A corresponding relation between the object’s parameters and the sensor output values is found by curved surface formulation. With the help of two dimension calibration and calibrated experiment

the coefficient in the regress equation under the condition of error-mean square is calculated in the light of minimum two multiplication principle. Thus

the object’s （parameters） can be calculated out by known coefficient’s multi-dimension regress equation once the sensor’s output value measured. When the algorithm was applied to piezoresistive pressure sensor

it not only （benefits） temperature compensation but also helps non-linear rectifying.The accuracy of sensor can be reached （to 0.05*#%）.Meanwhile

the sensor’s self-RS232 interface is applied to on-line compensation and rectification.