||The UK construction industry kills some 80 employees per year. This corresponds to a fatal accident rate of roughly 4.4 per 100,000 employees. Furthermore, major injury and 3-day injury rates are approximately 400 and 900 per 100,000 employees, respectively. Figures for the rest of Europe are similar, if not worse. It has long been realised that the reduction of hazardous events is fundamental to good construction safety management because these events have the potential to cause accidents, which may result in injuries and fatalities. However, there have been examples within the industry where hazard identification and the subsequent assessment of risk have been carried out by people ill equipped to identify all the hazards, assess their risks and suggest appropriate responses.Our research aim is to develop a knowledge-based system to aid in site safety management. The system will consist of a centralised database containing the combined knowledge and experience of all personnel within the company. This database is accessed via a user interface, which takes the form of a dynamic data-driven website and consists of four main applications that focus on the main areas of site safety management. The first application concerns creating and maintaining a company risk log, which can be used to identify hazards, assess risk, establish adequate hazard responses and report risk reduction performance. The second application is intended to aid in the method statement preparation process. Hazard referencing to tasks within the methodology and assigning significance values to tasks based upon assessed risk are the main features of this application, which should improve the level of hazard identification and enhance safe systems of work on site. Our system relies heavily upon historical data to provide an objective and dynamic evaluation of risk, rather than current subjective and static estimations of risk using the traditional method of risk matrices. Thus the other two applications are concerned mainly with entering data from accident reports and site safety tours into the central database. The other function of these applications will be to perform detailed analysis of accident causes, which will help safety managers to better respond to hazards to prevent future accidents.We are currently at the stage of developing a prototype version of the system. Field trials will be conducted between February and May for validation of the prototype. Validation will take the form of analysing method statements and risk assessments before and after implementation if the system to determine its effect on improving hazard identification, the assessment of risk, hazard response and accident rates. Qualitative evaluation will also be carried out. Questionnaires to, and interviews with, safety managers will give an indication of the usefulness of the system from a management and operational perspective. Doing the same thing with site foremen and operatives will allow us to determine the effectiveness of the outputs of the system, i.e. method statements and risk assessments, in enhancing safe systems of work.