Besides many advantages which cloud computing creates, there are different concerns such as security. In this paper, the conceptual model based on Biological Immune System (BIS) will be proposed in order to create security in cloud computing. BIS has several features such as distributed computing, self organizing, self-learning that are considered in distributed environments like clouds. In proposed model, five groups of autonomous agents are used. The structure of agents is based on Biological Agents (BA) that have memory and could use previous experiments. Agents have the ability to learn and interact with each other. Each agent has different functions. Their designs are based on B and T lymphocyte. This model is designed at two levels. In the first level, roles, relationship and activities of each agent are described and their components such as modules, programs and functions are shown in the second level. By using these intelligent autonomous agents, attacks can be identified and intrusion can be prevented. The proposed model is based on collaboration of the agents and doesn't need centralized management.