Optical assessment of Central Receiver Solar Tower systems should be based on annual simulations due to the high variability of solar radiation throughout the year. The most basic approach is using a discretization of the temporal domain which requires a large number of computationally costly optical simulations. This study proposes an alternative approach based on a discretization of the sky which reduces the required number of simulations and therefore the computational effort significantly. For the subsequent annual performance assessment data three different interpolation methods are presented, compared and their respective advantages and drawbacks are discussed. The methodology is demonstrated by means of three representative heliostat field setups: the PS10 field, the Gemasolar field and a small, generic field. In this context, the dependency of the interpolation accuracy with respect to the resolution of the sky grid is investigated. For this purpose, annual optical efficiency weighted with direct normal irradiation is used as an integral figure of merit, while the Root-mean-square deviation is calculated as a benchmark value for single time step assessment. Eventually, it is shown that the proposed method decreases the required time for the annual optical assessment by several orders of magnitude while maintaining the original accuracy.