Objective: To explore the change of pathology and the clinical response rate treated by neoadjuvant chemotherapy with MVP regimen for non-small cell lung cancer. Methods: This is a randomized study in patients with stage I-lIIa. Among them, 46 patients enrolled in neoadjuvant chemotherapy treated by 1-2 course MVP regimen. MMC was given 6 mg/M2 by intravenous (I.V.) infusion on day1, VDS 2.5-3 mg/M2 I.V. on day1, 8 and/or day15, DDP 90 mg/M2 I.V. on day1. The treatment was recycled every 28 days. The clinical RR evaluated with WHO criteria. All surgical samples were classified with pathology. Results: The overall response rate in 2 courses chemotherapy is better than that in 1 course (P<0.01). The number of patient with pathology grade I-II in 2 course chemotherapy is higher than that in 1 course (P<0.01). But the RR can not completely translated into pathology grade I-II. The pathology grade I-II is closely related with tumor involvement (T) (P<0.01) but not closely related with regional lymph node metastasis (N). It is reasonable to use RR together with PCR to judge the chemotherapy response. NR patients can not be regard as chemotherapy failure. No serve toxicities and surgical mortality were observed. Conclusion: MVP regimen is an effective neoadjuvant treatment regimen for I-IIIa NSCLC.