Objective To investigate the relationship between smoking and lung cancer by evidence-based evaluation. Methods Using Meta-analysis method, the results of 29 case-control studies involving the relationship between smoking and lung cancer in recent decade were analyzed by Review Manager 4. 2 software. Results The association between smoking and lung cancer was significant ( Z =12. 16, P lt; 0. 000 01) , and the pooled OR value was 5. 75( 4. 34, 7. 62) . The population attributable risk percentage( PARP) of smoking was 69. 16% . The pooled OR of 1-10 cpd( cigarettes per day) , 10-20 cpd, 20-40 cpd and more than 40 cpd were 1. 97( 1. 69, 2. 30) , 5. 20( 3. 54, 7. 62) , 7. 46( 5. 22, 10. 67) and 15. 14 ( 5. 27, 43. 44) respectively. The pooled OR of less than 20 years of smoking duration, 20-40 years and more than 40 years were 1. 25( 1. 01, 1. 53) , 5. 10( 3. 03, 8. 57) and 10. 77( 7. 30, 15. 89) respectively. While the pooled ORof less than 10 pack-years, 10-20 pack-years, 20-40 pack-years and more than 40 pack-years were 1. 73( 1. 01, 2. 96) , 3. 73 ( 3. 02, 4. 61) , 5. 69 ( 3. 79, 8. 54) and 8. 41 ( 4. 56, 15. 51) respectively. The pooled OR of initial smoking age less than 15 years old, 15-20 years old and more than 20 years old were 13. 31( 7. 09, 24. 97) , 7. 21( 4. 51, 11. 52) and 4. 74( 3. 47, 6. 47) respectively. The pooled OR of quitting smoking for 1-10 years, 10-20 years and more than 20 years were 7. 16( 4. 70, 10. 91) , 2. 12( 1. 16, 3. 86)and 1. 47 ( 0. 67, 3. 20 ) respectively, and more than 20 years of quitting smoking had no significant difference. The pooled OR of light smoking and deep smoking were 3. 26( 1. 24, 8. 58) and 8. 07( 4. 67, 13. 94) respectively. Conclusions Smoking is an important risk factor of lung cancer. Meta-anlalysis revealed cigarettes comsuption per day, smoking duration, total amount of cigarettes ( pack-years) , smoking behaviour( depth) , initial age of smoking and duration of quitting smoking can increase the risk of lung cancer.