In the context of climate change and agrosystem complexification, process-based models of the reproductive phenology of perennial grasses are essential to optimize the agronomic and ecologic services provided by grasslands. We present a functional–structural model called L-GrassF, which integrates the vegetative and reproductive development of individual Lolium perenne plants. The vegetative development in L-GrassF was adapted from a previous model of perennial ryegrass where leaf elongation and tillering dynamics partially result from self-regulated processes. Significant improvements have been made to this vegetative module in order to deal with the whole growing cycle during which plants are exposed to contrasting temperatures. The reproductive module is a new functionality describing the floral induction of the individual tiller from daily temperature and photoperiod as well as its phenological state. From the interactions between the vegetative and reproductive developments, L-GrassF simulates the dynamics of plant architecture, the floral transition and heading date (HD) at tiller level. A sensitivity analysis was performed on L-GrassF and showed that parameters controlling the kinetics of leaf elongation and leaf appearance rate have a significant impact on HD. After calibration, L-GrassF was able to simulate the HD on seven L. perenne cultivars grown in a broad range of environmental conditions, as provided by an independent data set. We conclude that L-GrassF is a significant step towards better prediction of grassland phenology in contrasted conditions.