FLuorescence or backscattered signals measured over a plant canopy after a picosecond laser shot (nadir viewing) are composed of elementary contributions coming from different illuminated leaves, branches, twigs, and the soil background. They are affected by time delays, depending on their levels within the canopy. The global signals measured are then rather complex and depend on canopy architecture and fluorescence characteristics. Also, an amplitude decorrelation exists between fluorescence and backscattered signals, due to the difference existing between reflectance and fluorescence properties of plant canopy components. For this reason it was necessary to develop a specific method for retrieving the mean fluorescence lifetime of a complex plant canopy under field condition. In a first step, the position and the reflectance characteristics of each reflecting canopy element is determined from the deconvolution of the backscattered signal. In a second step, the parameters obtained are introduced into the fluorescence function in order to determine the mean fluorescence lifetime. The validity of the deconvolution method has been tested for numerous situations, using simulated laser shots on a 3-D canopy mockup. The results obtained show a very good agreement between estimated and input parameters.