In the present study, a six-directional sampling method (SDSM) and six-directional sampling mean mixing indices (SDS-mMI) i.e., SDS-mLMI, SDS-mRSD, and SDS-mSMI based on SDSM are proposed for the first time by considering Lacey mixing index (LMI), relative standard deviation (RSD) and sub-domain mixing index (SMI) as basis to assess the solids blending performance in DEM simulations. The proposed approach consists of dividing the blender into small grids, and accumulating the mass and mass contributions of each type of particle by moving in six different directions to generate predetermined number of samples, finding the mixing index in each direction, and finding the mean value of the index over all the six directions. The proposed approach is evaluated for two typical blenders with cohesive/cohesion-less, mono/bi-dispersed particles systems. The proposed mixing indices based on the sampling approach are compared with the same mixing indices based on BCSM, and also with other non-sampling indices. Display omitted •Six-directional sampling method (SDSM) & mean mixing indices (SDS-mMI) are proposed.•SDSM applied and SDS-mMI computed for 50 L industrial bin-blender &14 L tote blender.•Cohesive & cohesion-less, mono & bi-dispersed particles considered as test cases.•SDS-mMI performance compared with BCSM, sub-domain, and non-sampling mixing indices.•Proposed SDSM based SDS-mRSD validated through experimental results from literature.