For any communication to be successful, trust is necessary. For inter-cloud communication, clouds interact with each other for resource sharing. Since they are unaware of their opposite party, there should be some mechanism by which the cloud gets an idea about them prior to the communication. This is accomplished through trust management systems that calculate the trust rating of clouds from opinions from their peers. There is no way to ensure whether these peers are genuine in their opinion or not. This paper proposes a method to reduce such problems by considering the latest history of communication of a particular cloud and ignore the opinions given by less trusted clouds.