Nonspherical ice crystal optical properties are of fundamental importance to atmospheric radiative transfer through an ice cloud and the remote sensing of its properties. In practice, the optical properties of individual ice crystals need to be integrated over particle size distributions to derive the bulk optical properties of ice clouds. Given a particle size distribution represented in terms of size bins, the conventional approach uses the microphysical and optical properties of ice crystals at the bin centers as approximations to the bin-averaged values. However, errors are incurred when the size bins are large. To reduce the potential errors, a kernel technique is utilized to calculate the bulk optical properties of ice clouds by computing the bin-averaged values instead of using the bin-center values. Comparisons between the solutions based on the conventional method and the kernel technique for different numbers of size bins from in situ measurements demonstrate that the results computed from the kernel technique are more accurate. The present study illustrates that, for a given size distribution, 40 or more size bins should be used to calculate the bulk optical properties of ice clouds by the conventional method. Although the accuracy of bulk-scattering properties can be improved by using fine bin resolutions in the single-scattering property computation, the advantage of using a precomputed database of scattering kernels allows efficient computation of ice cloud bulk optical properties without losing the accuracy.