Abstract: Combined with the definition of fire risk assessment, a clear explanation of the concept and development of urban area fire risk assessment is made. Aiming at the problem of regional fire risk assessment, it is proposed to divide the assessment methods into two categories: traditional assessment methods based on indicators and modern assessment methods based on machine learning. The evaluation process of two methods are summarized. According to the evaluation process, the common evaluation methods are classified and summarized. Combined with the current domestic and international status of regional fire risk assessment, this paper compares and summarizes the characteristics and scope of application of the two types of assessment methods, and proposes possible future development directions.

Key words: regional fire risk, risk evaluation method, traditional index evaluation, machine learning evaluation