Methane is widely used in chemical industries and for electricity generation. It is highly flammable even in low concentrations when mixed with other chemicals—it is also an asphyxiant as it will displace oxygen. Methane is a potential severe explosion risk and, as a landfill gas, should be effectively monitored and analyzed. Various researchers have proposed methods based on exact analytical approaches based on Kalman, H-infinity filter, etc., to estimate the trajectory profile of methane. However, existing methods cannot fully deal with non-Gaussian uncertainties that arise in the plume generation. A hybrid approach of ensemble forecasting and an H-infinity filter is proposed to provide better estimation accuracy. The proposed method can also facilitate sensor fusion to maximize the accuracy of estimation at the locations where sensors are not located.