Introduction - Weather parameters are the main limited factors for simulating yield for the crop growth models. It is well known the effect of single weather parameters on crop development but the interaction remains uncertain in condition of high variability of climate such as West Africa. This region is characterized by a continuous increase of temperature and a high uncertainty in the projections of the precipitation change. Nevertheless, in the particular case of Senegal mean precipitation is expected to decrease in the future. This study analyses the sensitivity of parameters: temperature, water and [CO2] on crop yield with a single response of each parameters on one hand and on the other hand a combination of parameters based on the response of temperature for each [CO2] level combine with six levels of water in each [CO2] x temperature combination. Materials and Methods - Two cereals: maize (improved maize and local maize) and pearl millet were simulated in two different soil types and two zones under different mineral and organic fertilizer applications. The simulations are performed with CELSIUS crop growth model couple with ANDERS economic model. The two models were calibrated and validated previously for maize and pearl millet in the given area (Ricome et al 2017). Two virtual experiment were setup with observed data as baseline from 1991 to 2010: (i) with no interaction for a total of 54 climate scenarios (11 Temperature levels + 8 water levels + 7 [CO2] levels + baseline) * 2 zones, (ii) with interaction (table1). Results and Discussion Preliminary results showed that improved maize had a higher response to the variability of climate with and without interaction in both zones irrespective of the soil type and fertilizer application. The response of the crop to the parameters considered is more uncertain under soil with low nitrogen content under the interaction case. Yield change remains constant under low fertilizer application compare to high fertiliz