Invasion of the oriental fruit fly, Bactrocera dorsalis, into new niches containing different food sources(a process referred to as host shift), may cause population genetic differentiation and sympatric speciation. To attempt to infer that experimentally, test populations were established by transferring a subset of the original populations, which had been grown on banana for many generations, onto navel orange, and then subculturing the navel orange population and banana population for at least 20 generations. Four pairs of SSR primers with high polymorphism on laboratory strains were used to detect population genetic differentiation. All six tested populations(the 5th, 10 th and 15 th generations of B. dorsalis fed on banana and navel orange, respectively) were found to have low genetic diversity. Furthermore, the genetic diversity of the navel orange populations was found to decline after being crossed for several generations. Populations initially were deviated from Hardy-Weinberg equilibrium, however, equilibrium was achieved with increasing numbers of generations in both of the host populations. Limited gene flows were found among the six populations. The Nei’s standard genetic distances between the two host populations of the same generation were initially low, but increased with generation number. Genetic distances between banana and navel orange populations of the same generation were lower than genetic distances between different generations grown on the same host plant. Analysis of molecular distance(AMOVA) results based on generation groups and host groups demonstrated that genetic variation among generations was greater than that between the two host populations. The results indicated that population genetic differentiation occurred after the host shift, albeit at low level. Biogeography and taxonomy of the B. dorsalis complex revealed that speciation of B. dorsalis might be tightly associated with host shift or host specialization of B. dorsalis following dispersal. Invasion of the oriental fruit fly, Bactrocera dorsalis, into new niches containing different food sources (a process referred to as host shift), may cause population genetic differentiation and sympatric speciation. To attempt to infer that experimentally, test populations were established by transferring a subset of the original populations, which had been grown on banana for many generations, onto navel orange, and then subculturing the navel orange population and banana population for at least 20 generations. Four pairs of SSR primers with high polymorphism on laboratory strains were used to detect population genetic differentiation. All six tested populations (the 5th, 10th and 15th generations of B. dorsalis fed on banana and navel orange, respectively) were found to have low genetic diversity. Furthermore, the genetic diversity of the navel orange populations was found to decline after be crossed for several generations. Populations initially were deviated from Hardy-Weinberg equilibrium, however, equilibrium was achieved with increasing numbers of generations in both of the host populations. Limited Generates were found among the six populations. The Nei’s standard genetic distances between the two host populations of the same generation were initially low, but increased with generation number. Genetic distances between banana and navel orange populations of the same generation were lower than genetic leaders between different generations grown on the same host plant. Analysis of molecular distance (AMOVA) results based on generation groups and host groups said that genetic variation among generations was greater than that between the two host populations. The results indicated that population genetic Differentially after the host shift, albeit at low level. Biogeography and taxonomy of the B. dorsalis complex revealed that speciation of B. dorsalis might be strongly associated with host shift or host specialization of B. dorsalis following dispersal.