Introduction: low birth weight (< 2,500 g) is the principal determining factor of infant and neonatal mortality. Objective: know the frequency of low birth weight in Peru, the associated risk factors and the strength of the association. Material and methods: a transversal study of 10,070 children born between 1991 and 1996 was performed. The sample included all the districts of the country and the urban and rural areas. The frequency of the risk factors (geographic, sociodemographic, economic and reproductive) and the association of these factors with the low birth weight were analyzed with a logistic regression model. Results: the national LBW incidence was 8.97% with 7.0% in the coast, 11.3% in the mountains and 12.4% in the forest. Using the multifactorial logistic regression analysis, the odds ratios (OR) were: height greater than 2000 m (Subregions of Quechuan, Sunni or Julca, highland region of the Andes and the Himalayas) (OR = 1.2) without education (OR = 2.3) or only primary education (OR = 1.5), without prenatal control (1.8) and without marital union (OR = 1.5). Conclusion: the strength of the association of the risk factors is modest but knowing them can be useful for low birth weight prevention programs, detecting those pregnant women at risk and paying greater attention to them.