Background and purpose: Early and accurate diagnosis of stroke by emergency medical service (EMS) paramedics is critical for reducing pre-hospital delays. The Los Angeles pre-hospital stroke screen (LAPSS) has been widely used as a validated screening tool for early identifying stroke patients by EMS paramedics. However, validation of LAPSS has never been performed in Chinese stroke population. This study is aimed to verify the LAPSS for early identifying stroke patients in a Chinese urban EMS. Methods: 76 paramedics of five urban first aid stations attached to Beijing 120 EMS were involved. The paramedics were trained by professionals to quickly screen patients based on LAPSS. Potential "target stroke" individuals who met the base LAPSS screen criteria were identified. Sensitivity and specificity analyses of the LAPSS were calculated. Results: From June 10, 2009 to June 10, 2010, paramedics transported a total of 50,220 patients. 1550 patients who met the baseline screen criteria were identified as the potential "target stroke" population. 1130 patients had the completed LAPSS information datasheet and 997 patients were clinically diagnosed with stroke. The average time of completing the LAPSS was 4.3 +/- 3.0 minutes (median, 5 minutes). The sensitivity and specificity of the LAPSS in this study was 78.44% and 90.22%, respectively. After adjusting for age factor by excluding patients of >45 years old, the sensitivity was significantly increased to 82.95% with specificity unchanged. Conclusion: The paramedics of Beijing 120 EMS could efficiently use LAPSS as a screening tool for early identifying stroke patients. While the sensitivity of LAPSS in Chinese urban patient population was lower than those reported in previous LAPSS validation studies, the specificity was consistent with these studies. After excluded the item of "Age >45 years", the sensitivity was improved.