On the basis of full study on the structure principle and temperature characteristics of silicon torsion pendulum capacitive accelerometer

temperature compensation model for micro-electro-mechanical system（ MEMS）accelerometer is constructed by back propagation neural network（ BPNN） optimized by mind evolutionary algorithm（ MEA）. Temperature compensation is realized in real-time. By temperature experiment

calculate temperature model parameters optimized

realize temperature compensation in realtime. Experimental results show that the temperature coefficient of scale factor

the full temperature zero bias range and the nonlinear are decreased from 438 × 10-6/℃

58. 8 mgnand 1 576 × 10-6 to 78 × 10-6℃

2. 7 mgnand 266 × 10-6 respectively after compensating with the temperature model. The experiment improves the temperature characteristics of the accelerometer greatly and proves the effectiveness of temperature compensation method.