The light-response curve of photosynthesis is an important tool used to study plant ecophysiology and can provide a scientific basis for the response of plant photosynthetic characteristics to environmental factors. At present, there are five common light-response models of photosynthesis. To gain deeper insight into the applicability of different light-response models of photosynthesis and the photosynthetic physiological characteristics of Populus euramericana cv. 'Zhonglin46', two typical light-response curves of photosynthesis in P. euramericana cv. 'Zhonglin46' leaves, one under drought stress and the other under control conditions, were measured using a CIRAS-2 portable photosynthesis system. The light-response data were divided into two groups: one set of data was used to fit light-response curves, and the other set of data was used to test them. The accuracy of the fitting and the predictions of the different models were evaluated by mean square error and mean absolute error. The results showed that the light-response curves of P. euramericana cv. 'Zhonglin46' under drought stress matched the light-saturated inhibition type and that those under the control condition matched the approaching light-saturation type. The two new models (i.e., the modified rectangular hyperbola model and modified exponential model) fit the two light-response curves and their characteristic parameters well, and the fitting results of the two models were similar. Conversely, the three traditional models (i.e., the rectangular hyperbola model, nonrectangular hyperbola model, and exponential model) did not fit the two light-response curves well; in particular, they overestimated the maximum net photosynthetic rate, underestimated the light saturation point (LSP), and did not fit the net photosynthetic rate during the light-saturated stage. The LSP calculated by the "linear method" combined with the traditional models was significantly lower than the measured values; additionally, the appropriate value of the proportional coefficient was difficult to determine, and the assumed value (empirical value) could easily lead to unreliable results by the "coefficient method". The "estimation method" based on the measured light-response data was still a relatively accurate, simple, and practical way to determine LSP. In addition, the nonrectangular hyperbolic model also had good accuracy and applicability in fitting the approaching light saturation curve on the basis of the "estimation method" to determine LSP.