Background Reduced exercise tolerance is an important clinical feature of chronic obstructive pulmonary disease (COPD) and is associated with poor prognosis. The 6-min walk test (6MWT) is widely used to assess exercise capacity; however, it is not commonly administered in primary medical institutions because it requires a suitable site and professional training. Ultrasound has great potential for evaluating skeletal muscle dimensions in COPD. However, whether skeletal muscle ultrasound can predict impaired exercise tolerance is unclear. Methods The study included 154 stable patients with COPD, who were randomly divided into a development set and a validation set. The thickness (RFthick) and cross-sectional area (RFcsa) of the rectus femoris were measured using ultrasound. Standardized RFthick (STD- RFthick) and Standardized RFcsa (STD-RFcsa) were obtained via standardization of RFthick and RFcsa by patients' height. Results Strong correlations were observed between the 6MWD and RFthick (r = 0.84, p < 0.001) and between the 6MWD and RFcsa (r = 0.81, p < 0.001). In the development set, the optimal cut-off values for men and women for predicting poor exercise tolerance were < 3.098 cm/m and < 3.319 cm/m for STD-RFthick and < 4.052 cm(2)/m and < 4.366 cm(2)/m for STD-RFcsa, respectively. In the validation set, the area under the curve (AUC) values for the prediction of a 6MWD < 350 by STD-RFthick and STD-RFcsa were 0.881 and 0.903, respectively. Finally, the predictive efficacy of STD-RFthick (AUC: 0.922), STD-RFcsa (AUC: 0.904), and the derived nomogram model (AUC: 0.98) for exercise tolerance was superior to that of the sit-to-stand test and traditional clinical features. Conclusions Rectus femoris ultrasound has potential clinical application to predict impaired exercise tolerance in patients with COPD.