Soybean meals are the most widely used source of high-quality vegetable protein for livestock, aquaculture, and especially for poultry feed. It may be processed in a variety of different ways to increase the protein concentrations and to deactivate anti-nutritional factors. The quality of soybean meal (SBM) is affected by nutrient and antinutritional content, as well as physical characteristics with one being particle size (PS). The PS is an important factor of feed manufacturing that impacts mixability, feed flowability, pellet quality, and animal performance. The objective of this study was to evaluate the effects of griding four different SBM types (Full-fat, FFSBM; Solvent- extracted SBM, SESBM; Expeller-extruded SBM, EESBM; and High-oleic SBM, HOFFSBM) at different roller mill gap settings on PS. These four sources of SBM were milled using a two-pair roller mill (Model C128889, RMS, Sea, SD) at 10 different settings: 0-0, 0-5, 15-10, 15-15, 20-20, 30-25, 35-20, 40-35, 50-25, and 50-50 (top pair-bottom roller pair settings). The PS of three replicate samples per roller-gap settings for each type of SBM were analyzed using the ANSI/ASAE S319.4 “Method of determining and expressing fineness of feed materials by sieving” standard using 13 sieves and an electric sieve shaker (ANSI) accompanied by 0.5 g of silicone dioxide dispersion agent and sieve agitators. The weight of particles retained by each sieve was expressed as a percentage of the total sample recovered. The average PS of the samples was expressed as the geometric mean diameter (dgw) and standard deviation (Sgw). Data were analyzed using one-way ANOVA and means were separated using Tukey’s test or student’s t-tests. Additionally, a regression analysis was conducted. These settings generated samples with average d gw of 574, 623, 681, 732, 778, 858, 881, 892, 905, and 947 μm, respectively. Regression analysis indicated negative quadratic effects (P<0.001) of roller gap on dgw with R2 of 0.83, 0.92 and 0.90 for FFSBM, HOFFSBM, and SESBM, respectively. These effects were observed in all the meals except EESBM which had (P<0.001) a linear effect (R2 = 0.79). The SESBM had the largest particle size followed by FFSBM, HOFFSBM, and EESBM. However, both FFSBM and HOFFSBM broke with similar dgw (P < 0.001). The Sgw decreased (P<0.001) linearly for EESBM (R2 = 0.73), FFSBM (R2 = 0.73), and HOFFSBM (R2 = 0.83) as the roller-gap was reduced. The Sgw was lower for the FFSBM and HOFFSBM than the SESBM and EESBM.