In order to forecast the survival of patients, we collected 417 cases of common epithelial carcinoma of the ovary as a population for this study, then selected 141 cases who had survived more than 3 years and 131 cases who had died within 3 years as a sample of statistical inference. The sample was computerized inputting 25 factors which were obtained mainly on the first surgery. We selected 8 factors with 46 items considering their weight to forecast the prognosis on convenience. The selection of factors was carried out by following three basic principles. First, we selected factors that retained high partial correlation coefficients. Second, we selected factors which did not overlap each other in their contents. Third, we selected factors which were not decided intentionally. Accordingly, the score for forecasting the 3-year survival of patients with ovarian carcinoma was framed. The correct discrimination rate of the score when it was judged at the 0 point was 86.8% in the sample group. However, in the sample group outside the population, it was 83.1%.