© 2020, African Association of Insect Scientists. The Mexican fruit fly, Anastrepha ludens (Loew) (Diptera: Tephritidae), is the most devastating pest of citrus in Mexico. Knowledge of the genetic relationships and structure of Mexican A. ludens populations is limited. The aim of this research was to characterize the genetic variability and structure of the Mexican fruit fly by using AFLP markers (three EcoRI/MseI primer combinations). The specimens were collected from three hosts, namely, Casimiroa edulis La Llave & Lex [white sapote], Casimiroa greggii (S. Watson) F. Chiang [yellow chapote], and Citrus X paradisi (Macfad.) [grapefruit], at four locations in the states of Veracruz (Teocelo and Jalapa), Nuevo León (El Jarro), and Tamaulipas (Troncones) in Mexico. A high polymorphism percentage (80.4%) allowed the differentiation of individuals from different A. ludens populations, and high intraspecies genetic variability (DI = 78.4%) was detected. The A. ludens populations clustered into two groups: group one included flies from Veracruz [grapefruit, Jalapa] and Nuevo León [yellow chapote, El Jarro], and group two included populations from Tamaulipas [yellow chapote, Troncones] and Veracruz [white sapote, Teocelo]. The populations from group one were from locations at lower altitudes (750 m above sea level), which had warmer (22 °C average temperature) and more humid (1470 mm of annual precipitation) climatic conditions than those of group two (1600 m, 18 °C and 965 mm). Both population structure and very high genetic differentiation (FST > 0.38) associated with geographical and/or host origin were found in A. ludens from Mexico. The genetic differentiation in Mexican A. ludens populations suggests reduced gene flow and reproductive isolation resulting either from habitat fragmentation and genetic drift or from natural selection for specific environmental (climate, geography, pest management strategies) or host (chemistry, phenology) condition. Genetic analysis could increase the success of Anastrepha control programs by improving knowledge of host-parasitoid population genetics, allowing matching of parasitoid species to their preferred pest populations.