The objective of this study is to evaluate the effect of Farmer Field Schools (FFSs) on cocoa yield in the Centre and South-west regions in Cameroon. The evaluation of FFSs is important for improving cocoa production by agricultural research and extension. This evaluation is done by using the “Propensity Score Matching” technique and Rosenbaum bounds sensitivity analysis method to ensure robustness of results. The data comes from IITA surveys conducted in 2009 and concerned 201 cocoa farmers in the Centre and South -West Regions in Cameroon for the 2008/2009 cocoa season. The sample cocoa farmers contain participants and non participants to FFSs. Results show that participate to FFSs have a positive and significant effect on the cocoa yield per hectare. This effect is estimated at about 97 kilograms per hectare. This effect is statistically significant at 10%, only for the kernel matching but not for the five nearest neighbors matching. This weak significant of FFS impact could be lead to the fact that farmers who joined FFS are those who have low capacity in cocoa production compare with those who did not joined FFS. This can also lead to the poor quality of FFSs training. The first recommendation is to promote the FFSs in the area where there is no FFSs by highlighting agricultural extension service. The second recommendation is to improve the quality of training in Farmer Fields Schools in other to increase their impacts on cocoa yield. In the current context of agricultural intensification challenges in developing countries, this analytical framework is of interest for policy makers for identifying conditions of farmer participation to FFSs and designing effective motivation strategies.