The proficiency of gross motor skills in early years is a foundation to the introduction of physical fitness and the development of more advanced and specific movement skills in later life (Bouchard, McPherson & Taylor, 1992; Gallahue & Ozmun, 1998). Therefore, a valid movement skill assessment tool is essential. It not only serves as a device for detecting developmental delays but also permits the collection of valuable data that can facilitate program planning. One of the frequently used movement skill assessments was the second edition of Test of Gross Motor Development (TGMD-2; Ulrich, 2000). It was a process-oriented assessment provided with criterion- and norm-referenced interpretations. Although the TGMD-2 was validated on US children, information regarding its reliability and validity was limited. Further, more analyses were needed to examine the structural validity of this testing instrument in countries other than US. The purpose of this study was to examine the factor structure of the TGMD-2 as applied to Chinese children. The TGMD-2 was administered to 1251 Hong Kong Chinese children (N = 1251) by a research team, which consisted of one investigator and four testers. After data screening, a total of 1228 cases (N = 1228) were used for further analysis. The data set was randomly split into two halves: one for exploratory factor analysis (EFA) (n = 614) and the other for confirmatory factor analysis (CFA) (n = 614). Using the criteria of scree test and Kaiser's criterion, a principal axis factor analysis with oblique rotation determined two factors, with five items (Hop, Slide, Dribble, Kick and Strike) loaded on both factors. These items were eliminated progressively and additional principal axis factor analyses were computed. The EFA results with eight variables (Hop, Gallop, Jump, Leap, Run, Catch, Throw and Roll) were the desirable EFA results: two interpretable factors were identified, with all items loaded on factors suggested by Ulrich (2000). The two-factor structure of the TGMD-2 was tested using confirmatory factor analysis with maximum likelihood estimation. Although the two-factor model of the TGMD-2 did not represent exact fit (x2 = 180.21, df = 53, p < .01), the x2/df ratio and goodness-of-fit indexes supported that the model was tenable (x2/df = 3.40, GFI = .95, RMSEA = .06, SRMR = .04, NNFI = .97, CFI = .97). The findings of this study suggested that the two-factor structure proposed by Ulrich (2000) fitted the data of Hong Kong Chinese children. Keyword(s): assessment, motor skills, research