The efficiency estimation and the interpretation of its behavior are of extreme interest for primary producer in agriculture as well as for policy makers. In Kosovo one of the main objectives of Agriculture and Rural Development Plan 2007-2013 and 2014-2020 is to improve competitiveness and the efficiency of primary agricultural producers and to attain sustainable land use. Regardless of this, there was a lack of studies on farm efficiency estimation and the productivity changes of the agriculture sector in Kosovo. Therefore, the conducted study of this thesis focuses on estimation and the analysis of efficiency at farm level. More specifically, the study aimed estimation of technical, economic, and environmental efficiency of the farms oriented on tomato, grape and apple production. In addition, identification of the factors that extensively explain the variation of the efficiency scores among farms was sought. The study was based entirely on primary data, collected in three different stages. In the first stage, a survey using structured questionnaire was conducted with 120 farms which were distributed equally for each selected production system in the study. Farm efficiency scores were obtained using a Data Envelopment Analysis, which is a linear programming optimization technique that measures relative efficiency of a set of comparable units. In general, the efficiency scores for three different production systems were high, showing that there was little space for efficiency improvement. On average, tomato farms tend to be more technical efficient, followed by scale, revenue, and cost allocative efficiency. Farmers oriented in grape production were very scale efficient, followed by technical, revenue and cost allocative efficiency. Apple farms on average were performing relatively well in terms of technical efficiency which was the highest on average, followed by revenue efficiency and scale efficiency. Factors which were proved to be statistically important in explaining the variation of the efficiency scores among the farms were household size, farm size and number of cultivated crops, number of land plots, farmer´s education and experience in farming. In terms of the position in ranking between technical and environmental efficiency estimation, three different groups of farms were found. The group of farms which showed increase in ranking at environmental efficiency when compared to the technical one. Farms with no difference in ranking, and a group of farms showing a decrease in ranking at environmental efficiency compared to the technical efficiency. Farms which displayed an increase in ranking were mostly farms that improved or maintained good quality of soil at farm land and good level of agro-biodiversity provision. The second group of farms showed no difference in ranking, as they were fully efficient in technical and environmental efficiency estimation. The third group of farms which showed a decrease in ranking were those farms performing weakly in both technical and environmental efficiency. This group of farms were also having lower soil quality at farm land and lower agro-biodiversity when compared to the averages of total sample.