Creativity plays a significant role in many fields, including design. However, evaluating creativity is a time-consuming and complex process that usually relies on human expertise. Recent advancements in computational techniques have brought the potential opportunity for using Artificial Intelligence (AI), especially the Large Language Models (LLMs), to assist designers in evaluating their creative ideas. This paper introduces an LLM-driven approach for evaluating creativity, focusing on novelty and usefulness, through a structured evaluation procedure. In addition, the Chain-of-Thought (CoT) technique was also adopted to enhance reasoning capabilities for LLMs in the proposed evaluation approach. Furthermore, a novel multi-agent structure was introduced where multiple LLMs acted as evaluators and analysts to collaborate for more reliable evaluation results. This paper also presents a detailed experiment, and the results show that LLMs can provide reliable evaluation results comparable to those of human experts. This paper is one of the first works that explores the use of LLMs to assist creativity evaluation in design and indicates the potential of developing LLM-based creativity support tools for human designers and researchers.