Life cycle assessment (LCA) has been the primary tool for the quantification and
comparison of the environmental impact of product systems. Despite ongoing
development of the LCA methodology, uncertainties are often not adequately
addressed in LCA research. The large variety of approaches to treat uncertainties and the lack of uniform guidelines prevent a widespread adoption of uncertainty analysis, while its importance is generally acknowledged. Especially in the agricultural sector, where LCA is often used as policy support, uncertainties are of great importance. This thesis aims to provide guidelines for practitioners of agricultural LCA in developing countries, as agriculture is often an important sector there. To this end four different method of uncertainty propagation are tested on a case study of sugarcane cultivation methods in Thailand. These methods are compared and evaluated on their suitability for agricultural LCA in developing countries, complemented by an expert survey in a Thai LCA research network. The results show that there is a general lack of knowledge among LCA practitioners on uncertainty analysis, while primary data is often abundant. Also, sampling methods can provide great insight into the output uncertainties, but are also complex and data intensive. Analytical or fuzzy interval methods could be a good alternative when knowledge or resources are lacking. LCA practitioners should be better guided in choosing and performing the most suitable propagation method for their situation. To this end, a decision tree for choosing the most suitable method depending on the user and the type of study closes this thesis.