There is a growing demand for the development of business applications composed of multiple components. Cloud computing address this by offering on-demand, scalable and cost-efficient access to resources. This paper proposes an automated approach to simplify Cloud-based business application development. The approach involves requirement elicitation, service discovery, composition plan generation, ranking, and deployment. It uses a Linked-USDL-based ontology to describe operational, technical, business, and semantic aspects of services, ensuring that all necessary service dependencies are considered. This approach minimises the number of providers, the user intervention and technical expertise, improving development efficiency. A prototype was built with Protégé and deployed via Docker. Evaluation involved three experiments using a virtual dataset of 120,000 cloud services. Results showed high functional similarity between user requirements and selected services, effective reduction in the number of providers when optimisation was applied, and improved user experience through automatic service composition outperforming platforms like Bitnami and Serchen.