Spectrum Sensing is an emerging technology in the field of wireless communication. It is an essential functionality of Cognitive Radio (CR) where it is used to detect whether there are primary users currently using the spectrum. Selection of suitable spectrum sensing technique is an important task, and it depends on accuracy and speed of estimation. Energy Detection technique is the most commonly used method for spectrum sensing. Non-cooperative spectrum sensing i.e. signal detection by single user suffers from several drawbacks. Thesedrawbacks include shadowing/fading and noise uncertainty of wireless channels. Hence, to overcome these disadvantages, a new methodology called Cooperative Spectrum Sensing (CSS) has been suggested in the literature. This thesis deals with the comparison of conventional spectrum sensing techniques and based on the computational complexity, accuracy and speed of the estimation, suitable sensing method i.e. energy detection technique will be selected. Here, we consider the optimization of conventional energy detection based CSS. In CSS, several CR's cooperatively detect the unused frequency slots called spectrum holes/white spaces. Generally, in CSS at the fusion centre, two data combining techniques are used which are softcombining and hard combining. Hard combining technique has gained importance due to its simplicity and it deals with three decision rules which are ‘AND rule', ‘OR rule' and ‘MAJORITY rule'. In hard combining only hypothesis output will be sent to the fusion centre, which decides about the presence of the primary user. For optimization, we have considered the network utility function and error probability. In order to achieve the goal we have proposed that the optimum voting rule is half voting rule also known as majority rule in ‘