Today, producer gas is being utilized as a fuel gas in boilers, internal combustion engines and turbines for heat and power generation. The composition of producer gas varies depending upon the gasification parameters. For improved process control and optimum utilization of these heat and power generating systems, it is desirable to monitor the producer gas composition in real-time. A new method and apparatus has been developed and lab-tested for quantitative characterization of producer gas. Spectroscopic and non-spectroscopic measurements are performed in order to detect both — spectrally active and inactive gases. Both methods are cross-sensitive to more than one gas. The measurements are then processed using multivariate statistical methods — principal components regression and partial least squares to fit a regression model which correlates the experimental measurements to the composition and heating value of producer gas. The fitted regression model is used to estimate the properties of unknown mixtures. The measurements and data processing are done in real time using a high speed hardware control and data acquisition system. A commercialized version of this sensor is expected to cost less than half the price of gas chromatographs, which are widely used in the gas industry today.