In the present work CPUE data of the Campeche shrimp fleet operating in the Gulf of Mexico were analyzed for the year 1997 using a Generalized Linear Model (GLM). Data were extracted from 240 logbooks of 171 shrimp vessels that summarized 33940 records of landing-catch. Additionally data included vessels characteristics (vessel name, overall length, horsepower, age built, gross tonnage, net tonnage, carrying capacity, hull material) landing location (Longitude and latitude), depth, fishing area and catchby specie. This information was integrated in amatrix including CPUE (Kg/effective day of fishing) as the dependent variable. A preliminary multiple regression analysis showed that net tonnage and carrying capacity were significant vessel characteristics. These variables plus landing locations, fishing area and depth were included as ranges codified in the GLM as follow: Shrimp monthly catches, five levels of carrying capacity (5-15 tons, 16-30 tons, 31-40 tons, 41-50 tons, > to 51 tons), three levels of net tonnage (39.5-70 tons, 71-90 tons, 91-110 tons), tbree levels of depth (< to 20 fathoms, 20-40 fathoms, 40-100 fathoms and 0-100 fathoms) three areas of fishing (1 = Regíon between Tamaulipas and north of Veracroz, 2 = Campeche Sound y 3 = Contoy zone). The GLM resultant was significant (p > 0.05), explaining a 42 % (R2 = 0.42) of the shrimp monthly catches. To carry out the estimations, August catches were used as reference since it was the month with the highest shrimp catch. This model allowed us to estimate the fishing power and fitted CPUE values by month, depth and carrying capacity. We could observe that CPUE changed with respect to spatial strata with the greatest on the regíons 1 and 2. These regíons are influenced directly by the fishing restricted periodo The highest catches were observed after restricted fishing period is opened.