Evaluation of systems, synthetic environments and human performance are generally complicated and time-consuming tasks. Existing evaluation systems are domain dependent and don't provide explanation on how the system reaches the evaluation results. Expertise is needed for evaluation process. Defining a model or a methodology, which simplifies and models the evaluation knowledge and speeds up the evaluation process, can obviously save cost, time and provide reusability. In this study, a knowledge representation of evaluation, which can also be used in distributed environments, was developed to handle the heuristic knowledge of experts from different domains and information from different sources for evaluation purposes. The evaluation knowledge was represented as reference model of evaluation objectives, production rules, evaluation measures, methods and parameters. Using of multi-agents is proposed for evaluation in distributed environments. Different types of multiagent organizations can be designed and developed. Multi-agents can simplify the evaluation process and reduce network traffic in distributed environments and decrease security problems among the network.