The Bayesian stable isotope mixing (MixSIAR) model was widely used in water source tracing. However, double isotope and single isotope calculated by the MixSIAR model led to different results in terms of previous studies. The effect of different nitrogen treatments on the water traceability results of wheat crops is still unclear. This study investigated the wheat root water uptake patterns at different nitrogen application gradients, as well as at different isotopes in the MixSIAR model. The results showed the main soil water uptake layer was relative constant at the treatment of 15% nitrogen reduction, and the water source was mainly from 0–10 cm soil layer from jointing to harvest. The main soil water uptake layer significantly varied at the treatment of 30% nitrogen reduction, which yielded the highest. 45% reduction of nitrogen treatment showed the deepest soil water uptake absorption on average with the whole wheat growth period. The consistency of the results between the two single isotope methods was better in the early stage of wheat and worse in the later stage. The inconsistency of root water uptake distribution results between the two single isotope methods might be influenced by the water transport mechanism of wheat stem and sampling errors. The double isotope method had the lowest uncertainty, but it might amplify the error of the sampling process. The uncertainty of the single hydrogen isotope method was less than single oxygen isotope method. This study provided a new evaluation of hydrogen-oxygen stable isotope traceability methods for wheat under different nitrogen treatments, and gave more ideas and insights for subsequent crop water traceability by the MixSIAR method.