Medical informatics application nowadays combines computer technology and clinical medicine to improve health care, medical education, and medical research. The availability of large amount and ever increasing of medical images produce a big crisis in existing processes for collecting patient's important data which need a great deal of many works to discover, select, and analyze the information. To reduce treatment delay and the errors when retrieving the data, resource broker technology must be bring into this field with providing proper indexing metadata. Existing medical imaging communities are increasingly aware of implementing broker in their technologies. In this paper, we present a solution to improve current medical image retrieval system by introducing goal-based cloud resource broker. Our solution is based on concept of goal which enables users to define required resources and to optimize the use of resources in the cloud. Several contributions have been found from this research where enhanced resources broker by using goal-based request have been proposed to maximize resources allocation within cloud environment and minimizing user involvement in resources discovery. © 2011 IEEE.