Prematurity, the major cause of perinatal morbidity and mortality, results from a multifactorial interaction of medical, historic, and psychosocial conditions. Although the literature contains several reports of prematurity risk-scoring systems, the relative importance of specific risk factors may depend on the population studied. This report represents the first prematurity risk-scoring system designed specifically for a predominantly Hispanic population in the United States. Retrospective analysis of 8240 births occurring at Harbor/UCLA Medical Center from July, 1979 to December, 1982 identified maternal prenatal risk factors that were found to be statistically related to prematurity. A linear logistic regression model was then employed to derive a composite risk score. Using the logistic risk scores, we developed a simplified model for identifying women at risk for preterm birth. The methodology and analyses provide a system for the development of population-specific risk scoring.