Most of the existing distributed data storage methods in wireless sensor networks (WSNs) rely on locating system of sensor nodes, which will lead to much energy consumption, and in particular, these methods are for the WSNs with uniform sensor distribution, and are clearly not applicable for the WSNs with nonuniform sensor distribution. To address this problem, this paper proposes a data storing algorithm based on the non-uniform node distribution, which aims at estimating the real distribution and the addresses of sensor nodes. To further save storage space and energy consumption, we integrate the Bloom filter with sensor nodes. The proposed algorithm offers an efficient search service so as to make data storage distribution and routing energy consumption more even, which can prolong network lifetime. The proposed algorithm also reduces redundant data with toleration of a given error bound to provide energy-efficient storage and further decreases the overhead of data routing and the waste of storage space.