Inter-row vegetative cover improves soil physicochemical properties and biodiversity but should not compete with main crop for water. Native grass species such as Leptochloa crinita (Benth.) Henrard var. californica provide forage value, require less water than conventional cover crops, and exhibit high tolerance to the edaphoclimatic conditions of arid regions. However, their small seed size hinders efficient handling and sowing. Seed coating increases seed size and weight, forming a pellet that improves handling, distribution, and protection. Nevertheless, limited information is available on coating techniques for small-seeded species. This study aims to optimize the coating process through a statistically designed experiment to evaluate key morphological and physiological seed traits. A Box-Behnken design was applied to optimize one compositional factor of the pellet and two operational parameters. Pellet traits such as size, heterogeneity, seed number, and residue were analyzed alongside germination and seedling development. Optimization was first carried out using edible amaranth seeds as a model organism and then validated with native grasses. The resulting models were statistically significant and allowed the identification of an optimal operational range. Germination was unaffected in single-seed pellets, and root growth was enhanced by the coating process. Overall, the proposed statistical methodology proved effective and broadly applicable for optimizing small-seed coating processes.