Energy harvesting (EH) is a highly promising technology which can support continuous power supply to energy-constrained communication system, such as machine-to-machine and sensor networks. Usually the main goal of the EH communication systems is to maximize the system through-put (i.e., the average achievable rate) under some causality constraints. The harvested energy can originate from the far-field radio frequency (RF) radiation of another device (e.g., a closely located communications transmitter which can transfer power and information simultaneously), or from the ambient RF radiation. If an energy-constrained system does not have dedicated communication resources (e.g., dedicated bandwidth), it can still be implemented as a secondary user (SU) system in a cognitive radio (CR) setup. The CR system utilizes one of the following three methods for spectrum sharing: interweave (i.e. opportunistic), underlay and overlay spectrum sharing. In the interweave approach the CR devices perform spectrum sensing to detect the unoccupied incumbent bands and then access the channel in this band. In the underlay approach, the CR devices operate in the same band as the primary users (PUs), and are allowed to interfere with the primary system under some predefined threshold. In this system PU and SU is located where both are requested for accessing the channel. PU uses licensed channel whereas SU uses unlicensed channel for transmitting the data. SU harvesting energy from ambient temperature while PU transmitted data via channel. SU starts to transmit the data when PU turns to silence. There are two approach proposed that is Non-Cooperation Mode and Cooperation Mode. Evaluate the performance by analyzing such as Optimal save ratio, optimal relay power and achievable throughput. EXISTING SYSTEM