A sewage pipeline harmful gas detecting device is designed based on sensor array and STM32 micro-coutroller unit（MCU） as microprocessor.By comparing three algorithms

such as least-squares regression

Gaussian process regression

and support vector regression algorithms

high-accuracy analysis algorithm is selected for different harmful gases.The detecting device achieves high-precision quantitative detection on harmful gases

and the concentration recognition error is controlled within a range of ±0.6 parts per million.