In the distributed computing paradigm, cloud computing provides a wide range of software, platforms and computing infrastructure. It has ability sharable elastic resource pools in dynamic. In this way, cloud computing offers many benefits. It reduces the cost of deployment in applications. It makes easy to increase the capacity of local infrastructures. And it can reduce power consumption by sharing resources. There are challenges over the benefits. The virtual instances needed to be shared equally distributed across different availability zones per demand. That can be categorized into operational and business challenges. In this paper, we implement the cloud prediction system using historical billing data and study the measurement of the model accuracy using Mean Absolute Error and Root Mean Squared Error. The trained model evaluates by the predicted and targeted value graph in this study. By leveraging historical financial data, this method aims to aid organizations in better foreseeing and managing AWS expenditures, leading to improved cost control and financial planning for cloud infrastructure.