Teachers’ attitudes toward teaching physical activity and fitness represent important factors affecting the quality of a school physical education program. However, there are few standardized instruments available that measure such attitudes, particularly among non-Western populations. This study assessed the validity and internal consistency of an attitude instrument with Chinese PE teachers. The “Teachers’ Attitudes Toward Curriculum in Physical Education” instrument was examined for use with Chinese PE teachers. The instrument includes 36 five point Likert items (1=Very important to 5=Not important) representing four domain areas: physical activity/fitness (PA), social development (SD), motor skill development (MS), and self-actualization (SA). Each domain area includes consists of nine items. The original instrument was first translated into Chinese. To control translation quality, a panel of five bilingulists in physical education were invited to evaluate item consistency between the English and Chinese versions of the instrument. A total of 330 PE teachers in Shanghai were then recruited. Because confirmatory factor analysis showed a “poor model fit” (AGFI=.74, RMSEA=.07, and CFI=.66), we used a principal component factor analysis with VARIMAX rotations to explore construct validity of the Chinese version of the instrument. Four factors were specified and the factor loadings to be ≥ .40. Based on the results, we dropped five items from the SA factor, three items from the PA factor, two items from the MS, and one item from the SD factor in the translated instrument. The deleted items revealed either cross loadings or factor loadings less than .40. We then performed a second principal component factor analysis with VARIMAX rotations for the remaining 25 items. The number of factors specified was again four. Results showed that the SA, PA, MS, and SD factors were retained. Factor loadings for all items exceeded .40 with no cross loading. The four factors accounted for 43.64% of the variance. Overall internal consistency of the instrument was acceptable with an alpha coefficient of .82. Alpha coefficients for subscales were .77, .76, .67, and .62 for the SD, MS, PA, and SA factors, respectively. Based on the obtained results, we now have an instrument that allows opportunities for cross-cultural research on teachers’ attitudes toward teaching physical activity and fitness. Keyword(s): assessment, international issues, measurement/evaluation