For alfalfa (Medicago sativa L.) to be profitable in the southeastern United States, enhanced persistence is required. However, studies assessing alfalfa's persistence in monoculture or mixtures in the region are limited. We aimed to determine the persistence and productivity of alfalfa and alfalfa–grass mixtures subjected to different harvest intervals. Three species combinations were established in 2015: alfalfa (A), alfalfa–tall fescue [Schenodorus arundinaceus (Schreb.) Dumort] (ATF), and alfalfa–bermudagrass [Cynodon dactylon (L.) Pers.] (AB). These were subjected to four harvest intervals (21, 28, 35, and 42 d) during the 2016, 2017, and 2018 growing seasons. Total forage accumulation (FA) was quantified during 2017 and 2018 and nutritive value and stem density were measured at the first and last harvests in those years. Forage accumulation decreased in all species combinations after 2 yr of management (A: 21 d, –63%; 28 d, –66%; 35 d, –50%; 42 d, –31%; AB: 21 d, –22%; 28 d, –59%; 35 d, –34%; 42 d, –19%; ATF: 21 d, –41%; 28 d, –62%; 35 d, –64%; 42 d, –41%), reflecting decreased alfalfa stem density, especially for A and AB. The differences in alfalfa stem density between the first and last harvests depended on the species. No differences were observed in ATF mixtures but for AB and A, longer harvest intervals had up to 90% fewer stems, thus decreasing alfalfa's persistence in the field. Lower stem density was observed for the 21‐d harvest interval; therefore, longer intervals could result in better field performance.