Currently the Oman economy depends on oil. Oil being a non-renewable resource, the neco eso government understands that oil deposits will be depleted someday. As a result, the government has put together vision 2020 strategies to diversify the economy away from oil to other non-oil sectors of the economy. Agriculture is one of the non-oil sectors. Under the 2020 vision agriculture is expected to grow by 4.1% annually so as to contribute 3.1% to the GDP of Oman by 2020. However, based on the recent past five year development plans, the performance of the agricultural sector has not been according to expectations. Among the challenges identified as being behind the poor performance of the sector include, low productivity of meat and milk from the local cow breeds. Assessing the efficiency of small scale dairy farms is therefore a first step towards finding solutions to the poor performance of the agricultural sector. This thesis therefore seeks to assess the efficiency of small scale dairy farms in Oman. Furthermore, the robustness of three popular functional forms namely, Cobb-Douglas (hereafter, CD), Translogarithmic (hereafter, TL) and Generalized Leontief (hereafter, GL) will be tested. Finally, regional efficiency differences (between the two main dairy producing regions of Salalah and Sohar) will be investigated. Result shows that the TL functional form is robust than CD and GL. Dairy farms in Oman are on average about 85% efficient, Regional wise, Salalah and Sohar dairy farms are respectively 81% and 77% technically efficient. These results imply that with better dairy husbandry practices, it is possible to increase milk production by 19% and 23% respectively in Salalah and Sohar with the same levels of inputs currently employed.