Diffuse proliferative lupus nephritis(DPLN) is the most severe form in lupus nephritis. We retrospectively analyzed 90 DPLN patients who were confirmed by kidney biopsy and treated at least for 12 months to compare the effectiveness of treatment modalities and to investigate the prognostic factors of DPLN. The patients were categorized as intravenous cyclophosphamide pulse(CY) group and oral corticosteroid(with or without cytotoxic drug; PO) group. When CY group(69pts) and PO group(21pts) were compared, there was no differences between two groups in sex, age, follow-up months, chronicity index(CI, 4.55 vs 3.76; CY vs PO, respectively), complications of treatment, mortality rate(4.3% vs 0%), remission rate of nephritis(at 3 year; 59.1% vs 75.5%) and renal survival rate(at 5 years; 88.0% vs 91.7%), b ut significant differences in frequency of nephrotic syndrome(66.7% vs 33.3%) and azotemia (30.4% vs 0%), activity index(7.14 vs 4.33) and relapse rate(2.9% vs 42.9%). When remission group (49pts) and nonremission group(41pts) were compared, the CI, initial azotemia, tubular atrophy, glomerular sclerosis, sex and duration of nephritis were prognostic factors for remission in univariate analysis, and CI(OR, 9.89) and sex(OR, 10.99) in multivariate analysis. When remission group(35pts) and nonremission(34pts) group were compared in CY group, the CI, initial azotemia and tubular atrophy were the prognostic factors in univariate analysis, and CI(OR, 7.69) in multivariate analysis. The initial azotemia was the only prognostic factor for renal survival. We concluded that the CI, sex and initial azotemia were the prognostic factors for renal remission and survival, but the treatment modality was not in DPLN. Therefore it is necessary to evaluate the prognostic factors before decision of treatment modality.