A microphysically based algorithm (named Snow Aggregation and Melting (SAM)) that models snowflakes composed of a collection of hexagonal columns by simulating a random aggregation process is presented. SAM combines together pristine columns with multiple dimensions to derive complex aggregates constrained to size‐mass relationship obtained by data collected from in situ measurements. The model also simulates the melting processes occurring for environmental temperatures above 0°C and thus define the mixed‐phase particles structure. The single‐scattering properties of the modeled snowflakes (dry and mixed phased) are computed by using a discrete dipole approximation (DDA) algorithm which allows to model irregularly shaped targets. In case of mixed‐phased particles, realistic radiative properties are obtained by assuming snow aggregates with a 10% of melted fraction.