Lameness in dairy cows is a clinical sign resulting from multiple diseases that affects animal welfare. The aim of this study was to evaluate the consistency of lameness prevalence estimations between farm managers and locomotion scoring conducted by trained observers, in confined and grazing dairy systems. The study was conducted on 18 dairy farms in Argentina. The locomotion of the lactating cows was scored by trained observers using a four-point visual scale from 0 to 3. Farm managers were interviewed about the number of lame cows in the herd. The consistency of lameness prevalence detected by the farm manager and the observers was assessed by computing the Lin's concordance and correlation coefficient. The comparison of grazing systems versus confined systems on lameness prevalence was analyzed using a generalized mixed model, assuming a binomial distribution for the errors. On average, farm managers estimated a lower prevalence of lameness ( p < 0.01) compared with the trained observers; 2.24% and 7.06%, respectively. Based on the estimations from trained observers, we could not detect differences ( p = 0.19) in lameness prevalence between confined and grazing systems.