Construction of a fuzzy inference system for describing the objective function in a fuel lattice composition optimization problem is proposed. This technique allows for incorporating human expertise in searching for the best radial fuel enrichment and gadolinia distributions in a typical boiling water reactor fuel lattice. The optimization procedure adopted is a modified tabu search algorithm. Evaluation parameters included in the objective function are obtained by the neutronic lattice simulator HELIOS. The performance of the new objective function is compared to the objective function consisting in a single sum of individual objectives pondered by weighting factors. Results show that the fuzzy inference system performs very well for modeling the objective function in order to qualify the investigated solutions in a fuel composition lattice optimization process based on tabu search. The best solution found is a lattice with the desired neutronic characteristic.