Hyperspectral (<1 cm-1) radiance measurements, continuous over a broad range of infrared wavelength, will be available in the near future from research and operational satellites. A novel technique has been developed to take full advantage of these measurements to improve the identification of cloud height and to retrieve the spectrum of effective cloud emissivity (the product of cloud amount and emissivity) simultaneously. This technique, called the Minimum Local Emissivity Variance (MLEV) algorithm, is based on the observation that the derivative of effective cloud emissivity with respect to cloud height depends on quantities that are highly variable spectrally. Because of this, errors in cloud height assignment will produce spectral dependent errors in the retrieved cloud emissivity. The correct cloud height, therefore, must be the one that results in the minimum local variance of the retrieved cloud emissivity, leading the name to the technique. Simulation of the algorithm, with proper treatment of observational noise and radiative transfer model errors, provides new insights into the algorithm. Initial verification using real data confirms the results from theoretical analysis and numerical simulation.