The main object  of this study is to examine  the  association  between vitamin  D and diabetes . national health interview survey (NHANES) dataset  has been  used  from 2009 to 2010 . the participants aged between  20 and 74 are  included  in the analysis. Logstic regression model  is used. Unadjusted and adjusted models controlling   by age, body mass index  ( BMI),serum cotinine  level , C-reactive protein (CRP),total cholesterol  (TC), hight-density lipoprotein (HDL) and Low-density lipoprotein (LDL)are estimated. All analyses are  stratified by  gender   to  examine  the  specific-gender  difference. In conclusion, there is  an  inverse  association  between vitamin  D   and  diabetes. Gender difference  have  been  shown significant  difference