Segregation distortion (SD) is often observed in plant populations; its presence can affect mapping and breeding applications. To investigate the prevalence of SD in diploid alfalfa (Medicago sativa L.), we developed two unrelated segregating F(1) populations and one F(2) population. We genotyped all populations with SSR markers and assessed SD at each locus in each population. The three maps were syntenic and largely colinear with the Medicago truncatula genome sequence. We found genotypic SD for 24 and 34% of markers in the F(1) populations and 68% of markers in the F(2) population; distorted markers were identified on every linkage group. The smaller percentage of genotypic SD in the F(1) populations could be because they were non-inbred and/or due to non-fully informative markers. For the F(2) population, 60 of 90 mapped markers were distorted, and they clustered into eight segregation distortion regions (SDR). Most SDR identified in the F(1) populations were also identified in the F(2) population. Genotypic SD was primarily due to zygotic rather than allelic distortion, suggesting zygotic not gametic selection is the main cause of SD. On the F(2) linkage map, distorted markers in all SDR except two showed heterozygote excess. The severe SD in the F(2) population likely biased genetic distances among markers and possibly also marker ordering and could affect QTL mapping of agronomic traits. To reduce the effects of SD and non-fully informative markers, we suggest constructing linkage maps and conducting QTL mapping in advanced generation populations.