Functional-structural plant models (FSPM) simulate plant development and growth, usually accompanied with visualization of the plant 3D architecture. GreenLab is a generic and mechanistic FSPM: various botanical architectures can be produced by its organogenesis model, and the plant growth is governed by the competition on biomass among growing organs. A distinguished feature of GreenLab model is that, its organogenesis (in terms of the number of organs) and growth (in terms of organ biomass) are formulated with recurrent equations. It facilitates analytical study of model behaviour, bug-proof of simulation software, and application of efficient optimization algorithm for parameter identification or optimal control problems. Currently several levels of GreenLab model exist: (1) the deterministic one (GL1): plants have a fixed pattern for development without feedback from the plant growth; (2) the stochastic level (GL2): pant organogenesis parameters are probabilistic; (3) the feedback model (GL3): the plant development is dependent on the dynamic relationship between biomass demand and supply (and in turn the environment). It makes it possible to deal with different kinds of behaviour observed in real plants. This paper presents recent typical GreenLab applications: (1) calibration of GL1 for getting sink and source functions of maize; (2) fitting of GL2 on organogenesis of wheat plant; (3) rebuilt of the rhythmic pattern of cucumber using GL3; (4) optimization of model parameters to improve yield, such as leaf (for tea) or wood quantity (for trees); (5) the possible introduction of genetic information in the model through detection of quantitative trait loci for the model parameters; (6) simulation of plant competition for light.