In this paper, we study power conservation techniques for multi-attribute queries in a wireless data broadcast environment. Most existing indexing techniques are based on a centralized tree structure and thus are inefficient for sequential-access wireless broadcast media. To conserve energy for mobile devices while maintaining acceptable data access latency, we extend the exponential index for single-attribute queries to multi-attribute queries. By maintaining a distributed structure and making full use of indexing space, the exponential index can reduce the energy consumption considerably. We conduct experiments to evaluate the performance of the extended exponential index against the well-known distributed tree index. The results show that the exponential index achieves a better performance than the index tree method.