Cloud computing is a new technology that has become a massive demand for computing solutions. Users can access computing resources from anywhere with the help of the cloud. Nowadays, more companies can hire cloud resources for storage and other computational purposes so that infrastructure costs are considered to be exceedingly reduced. The resource allocation mechanism based on the reservation method can be improved with a useful forecasting model that can provide almost well for developing resource allocation strategies. This can help to improve customer needs and the growth of businesses that are based on cloud computing. Service providers shall endure the cost of resources reserved in the cloud. Resources are allocated to reduce the associated costs. There are many algorithms implemented to predict the allocated resources in the cloud and to reduce the cost. This paper presents a Kalman filter with a Neuro-fuzzy system composed of an ANFIS optimized by the Quantum Genetic Algorithm. The algorithm was evaluated with actual cluster tracking data in Google and demonstrated the comparison method's weakness by showing much improved and better predictions.