In this study, Data Envelopment Analysis (DEA) was used to rank 25 onion producing provinces of Iran according to efficiency scores. The results illustrated that this approach could only divide the provinces into relatively inefficient (East Azerbaijan, West Azerbaijan, Razavi Khorasan, Khozestan, Fars, Kohgiluye & Boyerahmad and Markazi) and efficient categories (the other provinces). DEA ability in ranking reduced by changing the assumption from Constant Return to Scale (CRS) to Variable Return to Scale (VRS) in input-orientated model. The output-orientated model produced some better results, especially under VRS assumption. In order to achieve a full efficiency ranking, Andersen & Petersen and Chiao-Ping et al. methods were applied. The evaluation of average efficiency rankings showed that the provinces of Ilam, South Khorasan, Golestan and Sistan and Baluchestan had the best ranks from first to forth and the provinces of Rasavi Khorasan, East Azerbaijan, Khozestan and Fars took the least places in ranking, respectively. The comparison between yield and efficiency ranking results showed that only South Kerman district and Hamedan, Kermanshah, Zanjan and Kohgiluye and Boyeramad provinces were in the same ranking groups.