Wireless Sensor Networks (W) have been used in many applications during recent decades. In these networks, sensors are distributed over the different points of an area in various layouts. Basically, wireless sensors are used in detecting and transferring information due to their battery limitations. Therefore, energy consumption is a vital topic in these networks, and lots of algorithms and protocols in order to reduce the amount of energy had been proposed. Designing low complex systems with low energy consumption which can increase the lifetime of a network, besides preparing a balance between the transferred data and costs in a network are the greatest motivations in these studies. In this regard, compressed sensing is one of the methodologies for evaluating the related challenges. In this study, different methods proposed in energy reduction are introduced and a new evaluation based on compressed sensing is discussed. First of all, energy efficiency methods and their standards are considered besides compressed sensing methods in wireless sensor networks. After that, the way of using compressed sensing based on spatio-temoral features between the nodes is studied. In fact, using these features can be helpful in clustering the nodes and selecting a cluster head as a responsible node of a cluster for energy the efficiency of a network and increase its lifetime. In this thesis, a new algorithm in compressed sensing based on random walking in wireless sensor networks is proposed and it is shown that mentioned algorithm is not dependent on the sort of correlation between the nodes and can reconstruct the signal with higher accuracy. Also, a theorem is mathematically is proved to show the special reconstruction values in reconstructing the proposed algorithm. Keywords : Wiereless sensor Networks (W), Compressed sensing, Spatio-temporal Networks, Lifetime of a network