A key issue in agriculture is maximizing crop efficiency while respecting natural environment and farms sustainability. We put our interest on an application of smart farming, proposing a generic way to identify the parameters of a growth stage prediction model. We propose an approach based on a parametric model for data aggregation thanks to the Choquet integral. The case study concerns an implemented prototype that, for a given culture and several input sources, provides farmers with better predictions of the growth stages. The experimental results are very encouraging, the predicted day remaining stable despite presence of noise and local errors.