ThispaperpresentsasolutionusingartificialneuralnetworkstoreducetheeffectsofthetemperatureandhumidityoftheenvironmentontheresultsoftheTGS2602sensorinmeasuring NH3gasconcentrations.TheTGS2602inparticularandtheMOX(MetalOxidebasedsensorsingeneralhavehighsensitivity,fastresponse,longerservicelife,wideroperatingtemperaturerange,lowcost,lowpowerconsumptionbuttheirmaindisadvantagesarethestrongaffectionbyhumiditylevelandtheenvironmentaltemperature.Thismakestheproblemofeliminating(orreducing)theinfluencefactorsveryimportant.Inthispaper,asystemwithgassensor,temperatureandhumiditysensorstomeasuretheenvironmentalconditionsandMLP(MultiLayerPerceptron)networkstocalibratethesensorreadingwillbepresented.Thesimulationresultswillshowtheaccuracyoftheproposedsolution.