This study has evaluated the relationships between several petrophysical characteristics and seismic attributes of a siliciclastic reservoir in the Campos Basin, Brazil. Very thin and nonreservoir sedimentary beds and cemented layers with random distribution are the main heterogeneities in this reservoir. Such heterogeneities are below seismic resolution but we show that some seismic attributes are useful to obtain a better estimate of the nonreservoir cumulative thickness in the productive stratigraphic unit. A better understanding of spatial distribution and thickness of nonreservoir facies is useful to build realistic net pay and heterogeneity maps and recoverable oil estimations.Using models with velocities based on the well logs and petrophysical analysis, we computed 10 seismic attributes from synthetic seismograms. Some of them showed a good correlation with nonreservoir facies thickness. The vertical distribution, thickness and amount of nonreservoir layers in the models were randomly generated. The analysis using seismic attributes computed from real seismic data and nonreservoir facies thickness derived from the drilled wells confirmed the relationships observed with synthetic data.The real seismic attributes were then used to constrain the estimates of nonreservoir facies spatial and cumulative thickness distributions through kriging with external drift (KED). Stochastic simulation was applied to quantify the estimation uncertainties. The generated risk maps allowed the understanding of the nonreservoir facies distribution.