GreenLab is a recurrent discrete-time functional-structural model of plant growth and architecture. A method is presented estimating its parameters: the model is fitted to plant morphological and architectural data observed at one point of time. Since GreenLab output variables (number, size and fresh mass of organs) implicitly and nonlinearly depend on the model parameters, the fitting problem is solved by minimizing a generalized least-squares criterion and by implementing an iterative procedure. Fitting is satisfactorily performed on unbranched plants (cotton, maize, sunflower) using real data. The method is extended to more complex plants (i.e. with branches): a preliminary test on a virtual tree shows that the fitting algorithm also applies to such structured plants.