RESUMEN Se realizo un estudio descriptivo, con base en registros prospectivos de la dieta, para identificar los alimentos en las dietas bajas y altas en grasa y las variables asociadas. Area urbana de Bucaramanga, Colombia, Sur America. 1998-2003. Ciento sesenta y siete sujetos entre 20 y 60 anos seleccionados aleatoriamente, contestaron formatos con variables biologicas y socioeconomicas. Ademas, registraron y pesaron durante siete dias su dieta. De acuerdo al consumo dos grupos fueron establecidos; dietas con aporte bajo y alto de grasa al total de la energia; 30% respectivamente. Los alimentos fuentes se identificaron, la frecuencia de consumo semanal y los tamanos de porcion fueron comparados entre los dos grupos con t de student . La asociacion entre las variables biologicas y socioeconomicas con las dos dietas fue establecida con regresion binomial. El 40,1% consume una dieta alta en grasa, sin diferencia por ano (p=0,955), sexo (p=0,974) o Indice de Masa Corporal (p>F=0,993). El consumo de kilocalorias/dia no es menor en la dietas bajas en grasa con respecto a las altas; 2942 IC (2969, 3588) versus 3278 IC (2619, 3266) (p=0,151). De los 370 alimentos identificados, 45 explican el 70% del consumo de grasa. Ni los alimentos, la frecuencia de consumo, las variables biologicas o socioeconomicas explican el consumo de una dieta alta en grasa. La dieta es monotona. El tamano de la porcion es la unica variable que explica la clasificacion de dieta baja y alta en grasa. Palabras clave : Grasa dietaria, Registro dietario, Ingesta de energia, Ingesta de grasa, Preferencias de alimentos, Colombia. ABSTRACT We carry out a descriptive study, with base in prospective record of the diet, to identify the foods in the low and high diets in fat and the associate variables. Urban area of Bucaramanga, Colombia. 1998-2003. Hundred sixty seven subjects among 20 and 60 years old selected aleatorily, they answered formats with biological and socioeconomic variables. Also, they registered and they weighed during seven days their diet. According to the consumption two groups they were established; diets with contribution under and high of fat to the total of the energy; 30% respectively. The foods sources were identified, the frequency of weekly consumption and the portion sizes were compared among the two groups with t student. The association among the biological and socioeconomic variables with the two diets was established with binomial regression. 40,1% consumes a high diet in fat, without difference per annum (p=0,955), sex (p=0,974) or Index of Corporal Mass (p>F=0,993). The energy/day consumption is not smaller in the low diets in fat with regard to the high diet in fat; 2942 IC (2969, 3588) versus 3278 IC (2619, 3266) (p=0,151). Of the 370 identified foods or recipes, 45 explain 70% of the consumption of fat. Neither the foods nor recipes, the consumption frequency, the biological or socioeconomic variables explain the consumption of a high diet in fat. The diet is monotonous. The size of the portion is the only variable that he/she explains the classification of low and high diet in fat. Key Words: Dietary fat, Dietary record, Energy intake, Fat intake, Preferences of foods, Colombia.