Iterative spectrally smooth temperature/emissivity separation algorithm is an efficient algorithm designed for temperature and emissivity separation of hyperspectral data. The key point of this algorithm lies in the introduction of smoothness index used to evaluate the smoothness of emissivity curve, then the residual of atmospheric absorption lines is evaluated using the defined smoothness index, which is the basis of the object temperature optimization. The optimal object temperature corresponds to the minimum smoothness index.With the derived optimal object temperature, emissivity can be calculated easily. In order to evaluate the effects of smoothness index on the accuracy of ISSTES algorithm, 157 emissivity curves from MODIS UCSB library are selected to generate simulated dataset. Based on the simulated calculation, an optimal smoothness index is given.