This paper contributes to the empirical evidence on the gender differences in agricultural productivity. Using detailed household and individual data from the Uganda LSMS-ISA (2009-10; 2010-11) we estimate the value of productivity of crops grown per acre of harvested land at the household level, based on the gender of the land manager. Results from the Tobit model with fixed effects confirm the findings of the existing literature: controlling also for socio-economic variables and plot characteristics (soil quality, topography, distance from the homestead), as well as for the use of inputs (both labour and other inputs than labour) female managed plots are less productive than plots managed by men. Better individual agricultural data disaggregated by gender may allow to better identify the reasons of such productivity gap.