Objective　This study seeks to use a meta-analytical approach to quantitatively assess the results of applying the problem-based learning（PBL） teaching model and the traditional lecture-based learning（LBL）teaching model to basic medical courses in undergraduate medical edu-cation. Methods　The CNKI and VIP databases were electronically searched to retrieve randomised controlled trial studies that examined the use of PBL methods for basic medical courses in under-graduate medical education. In these studies PBL teaching model was used in experiment group and LBL teaching model was used in control group. Pass rate or test scores was used to evaluate the effect of learning. The retrieved documents ranged from the time that each database was first constructed to December 2012. After two researchers performed literature screening independently， data extraction， and quality assessment procedures in strict accordance with the inclusion and exclusion criteria of this study， the Stata 11.0 software package was utilized to conduct the meta-analysis. Results　This meta-analysis examined 28 studies that included a total of 3703 subjects. The modified Jadad scores of 20 studies（71.4%） were less than 4， and those of 8 studies（28.6%） were more than or equal to 4. The meta-analysis revealed that compared with the traditional teaching model， the PBL teaching model did not produce improved examination passing rates for a course [relative risk（RR）： 1.05， 95% confidence interval（CI）： （0.99， 1.10）， P=0.098] but could improve examination scores for a course [standardized mean difference（SMD）： 0.73， 95% CI： （0.51， 0.96）， P<0.001]. Conclusion　For basic medical courses in undergraduate medical education， compared to LBL teaching， PBL teaching can improve students' test scores but cannot improve students' pass rate. PBL used in basic medical courses has di-versified forms and lacks unified criterion， so there is still a long way to go for the appli-cation of PBL.