The goal of this study was to identify risk factors for diabetic peripheral sensory neuropathy in type 2 diabetes mellitus in a Chinese population. Peripheral sensory neuropathy was detected by quantitative sensory testing (5.07/10 g monofilament, neurometer and 128-Hz Riedel Seiffert graduated tuning fork). Those who had two or more abnormal quantitative sensory testings were defined as having diabetic sensory neuropathy. Of the 558 non-insulin dependent diabetes mellitits subjects, 62 (11.1%) had peripheral neuropathy. In 59 (10.6%) detection was by monofilament testing, 45 (8.1%) by graduated tuning fork, and 189 (33.9%) by neurometer. In a multivariate logistic regression model, age and insulin therapy were significantly associated with peripheral neuropathy. Age, serum triglyceride, height, and fasting plasma glucose were independently associated with large fiber neuropathy. Our results confirm the previously identified multiple risk factors of diabetic neuropathy. Different quantitative sensory testings detect different nerve fiber defects. The weak correlation between these tests indicates the need to use more than one test in screening for diabetic neuropathy.