In view of unfavorable situation that output precision of isopropyl alcohol gas sensor is often affected by humidity of environment

an improved BP neural network combined with genetic algorithm（GA）is proposed to compensate.The GA is used to optimize the initial weight threshold of the BP network to realize the function of compensating for different humidity of the gas sensor. Simulate the improved algorithm and compare the compensation effects of different algorithms. The experimental results show that the GA optimizes the BP neural network humidity compensation effect is better

compared with the standard BP network compensation error is reduced by 65.6 %

compared to the improved BP algorithm compensation error is reduced by 35.3 %

and the root mean square error is more small. The GA-BP network compensates the humidity of the isopropyl alcohol sensor

which makes the sensor have higher detection accuracy and humidity reliability.