Abstract: In cognitive radio systems, the information of primary user location and channel usage state, which can be obtained by spectrum sensing, dipicts the spectrum usage in spatial and temporal domain, respectively. And the accuracy of the information effects the efficiency of opportunity spectrum accessing of secondary users. Joint spatial-temporal spectrum sensing problem was investigated and modeled as a combined estimation problem of primary user location and channel usage state. Then, a joint spatial-temporal cooperative spectrum sensing method based on particle filtering was proposed. The main idea of the method is to estimate primary user location and channel usage state by their joint posterior probability density function that approximated by discrete random measure. Further, considering the estimated information in spatial and temporal domain, the proposed method adaptively selects those secondary users, who are closed to the primary user and symmetrically distributed in spatial domain, for cooperative spectrum sensing. The adaptive selection of secondary user improves the availability of the information provided by secondary users, resulting in a more accuracy of the estimation. Finally, the efficiency of this method was evaluated through simulations.