Ambient measurement and microenvironmental modeling were compared with personal measurement in Beijing, China to evaluate their capacity to determine personal exposure to PM(2.5). The comparison showed the association was insignificant between ambient and personal concentrations, but was significant between modeled and personal concentrations. The association between ambient and personal concentrations was improved for non-smoking dormitories, on heavily polluted days and on weekdays. The median difference was 41% between ambient and personal concentrations and 17% between modeled and personal concentrations. The factors affecting the association and agreement between methods were indoor sources and ubiquitous "personal cloud".