This document describes a PIC microcontroller and PC-based system using multiple gas sensors and artificial intelligence techniques for gas identification. Five commercial gas sensors are used to detect methane, carbon monoxide, and LPG at different concentrations and temperatures. The microcontroller collects analog voltage output from the sensors. Artificial neural networks are trained on the sensor data to identify gases based on patterns in responses to varying parameters like concentration, temperature, and load resistance. Experimental results show the sensors have different sensitivities to different gases and temperatures. The neural network approach can accurately predict gas concentrations online based on the sensor behavior patterns extracted during experiments. This system improves gas detection sensitivity and selectivity with high accuracy.