Our objectives were to determine whether the feedlot-level use of a direct-fed microbial (DFM; Lactobacillus animalis LA51 and Propionibacterium freudenreichii PF24; Bovamine Defend®, 2 × 109 CFU/g) was associated with fecal prevalence and concentration of E. coli O157:H7, and determine pen- and feedlot-level risk factors associated with fecal E. coli O157:H7 prevalence in cattle pens from commercial feedlot operations. Twenty commercial feedlots in Nebraska, ten that included DFM (DFM) and ten that did not (no-DFM), were sampled during the summer of 2017. In each sampling month, 22 pen-floor fecal samples were collected from three pens in each feedlot. Samples were subjected to cultural and molecular procedures for the detection of E. coli O157:H7 (immunomagnetic separation, plating on selective media, followed by PCR confirmation) and spiral plating for quantification. A total of 1,320 samples from 180 pens of finishing cattle belonging to 20 feedlots, which were sampled three times throughout a 12-week period, were processed and tested. Across all feedlots and sampling months, the mean within-pen prevalence was 13.5% (95% CI = 2.6–47.4%). The association between DFM status and the within-pen prevalence of E. coli O157:H7 depended significantly (p < 0.05) on the sampling month. The second sampling month between late July and mid-August corresponded to the highest within-pen prevalence estimates reported in this study, with no-DFM pens having a higher prevalence than DFM pens. After accounting for the DFM status, and based on multivariable analyses, sampling month, average pen body weight, and weather conditions were significantly associated with the within-pen fecal prevalence of E. coli O157:H7. Collectively, these findings demonstrate that the use of a DFM containing Lactobacillus animalis LA51 and Propionibacterium freudenreichii PF26 in feedlots showed potential in reducing fecal E. coli O157:H7 prevalence in cattle during times when prevalence peaks.