To analyze the factors affecting the quality of cardiopulmonary resuscitation (CPR) performed by medical staff in hospital and to explore the training methods to enhance their in-hospital emergency response capabilities. A cross-sectional study was conducted, involving medical staff of intensive care unit (ICU) and general internal medicine wards in China-Japan Friendship Hospital in December 2021. The American Heart Association (AHA) resuscitation quality improvement (RQI) model was used to evaluate the skills of the subjects in performing external chest compressions and bag-mask ventilation on adult and infant simulators. While ICU subjects were undergoing RQI model objective assessment, two instructors also provided subjective scoring for their operations. The study compared the differences in RQI model objective assessment scores between ICU and general internal medicine ward subjects, between doctors and nurses, in the RQI model objective scoring for adult and infant resuscitation, in the scoring differences of different positions for chest compressions, and the differences between traditional subjective scoring and RQI objective scoring when ICU subjects were assessed for compression and ventilation. A total of 75 medical staffs were enrolled, consisting of 50 from the ICU (including 24 doctors and 26 nurses) and 25 from the general internal medicine wards (including 10 doctors and 15 nurses). The ICU medical staff's scores for adult resuscitation skills were significantly higher than those of the general internal medicine ward medical staff [adult compression score: 82.5 (66.0, 96.5) vs. 65.0 (52.5, 74.5), adult ventilation score: 82.0 (68.8, 98.0) vs. 61.0 (48.0, 82.0), both P < 0.01]. The nursing group's compression scores for both adult and infant were significantly higher than those of the doctor group [adult compression score: 77.0 (68.5, 89.5) vs. 63.0 (40.8, 90.3), infant compression score: 54.4±25.1 vs. 41.5±18.5, both P < 0.05]. The compression and ventilation scores for the infant were significantly lower than those for adult resuscitation [compression score: 48 (29, 65) vs. 76 (58, 90), ventilation score: 56 (42, 75) vs. 76 (60, 96), both P < 0.01]. When the rescuer was positioned on the right side of the model, the compression score for the adult significantly increased [79.0 (65.0, 92.0) vs. 65.0 (51.3, 77.0), P < 0.05]. The ICU medical staff's traditional subjective scores of compression and ventilation assessments for adult were significantly higher than the RQI model objective scores [adult compression score: 88.8 (79.4, 92.5) vs. 82.5 (66.0, 95.5), adult ventilation score: 95.0 (80.0, 98.1) vs. 82.0 (68.8, 98.0), both P < 0.01]. Rich experience in emergency rescue is related to the improvement of CPR skills, and performing chest compressions from the right side of the adult model is more effective. Objective scoring of resuscitation skills based on the RQI model may more accurately reflect the performance of the trainees.