Objective This study aimed to evaluate the capacity of domestic laboratories for the detection of Salmonella in milk powder.Methods Artificially contaminated samples in this proficiency testing program were used. The samples were prepared by adding the bacterial enrichment solution to the matrix solution and then mixed， divided， and freeze-dried. The homogeneity and stability of the samples were tested. Three groups of proficiency testing samples were designed and two samples were randomly sent to each laboratory. The result was compared to the specified value to evaluate the performance of the laboratories that participated.Results A total of 337 laboratories in our country participated in the proficiency testing program， of which 323 were satisfactory， with the satisfaction rate at 95.8%； 14 laboratories had one or two unsatisfactory results. A total of 674 samples were sent to the laboratory， of which 658 had satisfactory results， with a satisfaction rate of 97.6%， 16 results were unsatisfactory accounting for 2.4%.Conclusion Different groups of samples were randomly numbered and randomly distributed， which played an essential role in preventing collusion. The samples prepared by the artificial pollution method fully considered the target bacteria， similar interference bacteria， and background bacteria， and were consistentwith the actual samples in terms of performance. The detection ability of Salmonella in the milk powder of national laboratories is generally satisfactory， and while a few laboratories need to be improved in terms of their detection capability.