Soybean is a candidate crop to increase the independency of Europe in leguminous protein crops. However, its adaptation to northern European regions is not yet well defined due to lack of long-term references. Herein, we simulated soybean yield potential in northern France and identified the main yield limiting factors under rainfed vs irrigated conditions. Two cultivars representing maturity groups 000 and 00 were planted within three different trials. Leaf area index, shoot and pod biomass, main phenological stages and yield were recorded to evaluate CROPGRO-soybean model predictability. Adjustment of genetic coefficients was performed prior to simulate yield on 21-years weather database (1999-2018) at Beauvais (France, N 49.46°, E 2.07°) and Estrées-Mons (France, N 49.88°, E 3.01°) under different water regimes and planting dates. Predictions showed that adding irrigation at grain filling period would increase yield potential to the level of non-water limited scenarios. Although simulated yield variability is reduced with irrigation, the remaining variability suggests that water is not the only yield-limiting factor. A tentative explanation is proposed by deriving environmental covariates from the model. The analysis confirmed the importance of precipitation amount (optimum around 200 mm) and duration (optimum around 60 days) of the flowering to physiological maturity period under rainfed conditions. Under irrigated conditions, increasing evapotranspiration and average minimum temperature affected simulated yield positively while increasing the number of days below 10°C had a negative impact. These results give insights for soybean crop management and bring indications to breeders for adapting the existing genetic material to northern Europe.