Facet theory was developed by Louis Guttman as a systematic approach for coordinating theory and research by means of formal a priori definitions of the domain of the study and of the variables that comprise the domain. And it serves as the basis for building a unified conceptual framework for the research. The main purpose of this study is to develop a new measurement instrument called teaching efficacy scale applying Facet Theory approach. Referencing to the findings of former studies and the interview results to 10 middle school teachers, a two-facet mapping sentence for the teaching efficacy concept was designed. Based on the structuples of the two facets 40 items for measurement of teaching efficacy were developed and a primary teaching efficacy scale (TES) was constructed. Based upon responses from 162 middle school teachers to the scale, using a combination of traditional reliability analysis and Smallest Space Analysis (SSA) 9 ambiguously located items in the SSA maps were removed and a formal 21-item teaching efficacy scale emerged. Using formal TES and the mapping sentence, hypotheses regarding the relations between the de?nitional framework and empirical observations were tested through smallest space analysis (SSA) and confirmatory factor analysis of data collected from a sample of 530 middle school teachers. The results demonstrate strong support for the de?nitional system and show a clear radax structure of teaching efficacy. The 2-dimension configurations of Facet A that was named tasks and Facet B that was named cognation resources and the 2-dimensional combination configuration of Facet A and Facet B had very small coefficients of alienation (<.14) suggesting that the configurations were very reliable. The region index of the configurations of Facet A, Facet B and the combination configuration of Facet A and Facet B are very large (>0.9) suggesting the facet definitions were strongly supported and the theoretical constructions were empirically confirmed. The items designed by Facet Theory approach empirically distinguish distinct character-based components of teaching efficacy. Facet A could be partitioned further according to the modality (knowledge transfer, organization and management of classroom and behavior and moral education of student) and Facet B could be partitioned further according to polarizing (the quality ability, student feedback, positive experience and emotionality). The combined 2-dimensional configuration of Facet A and Facet B turn out to be a radex presentation. By using the definitional mapping sentence of Facet Theory and SSA, the multidimensional theory and structural proposal was formed which was further tested by Confirmatory Factory Analysis (CFA). Obtained fit indices indicated a good fit between the data and the hypothesized factor structure. All parameter estimates were acceptable. Four conclusions could be summarized as: 1 Teaching Efficacy Scale can be defined from two facets. Facet A (task) consisted of three elements corresponding three factors and Facet B (cognitive source) consisted of four elements corresponding four factors. 2 The newly developed TES with two facets has high reliability and validity, which can be used as an evident instrument to measure teaching efficacy. 3 The teachers of secondary school concern more about students’ knowledge. For the evaluation of the teaching effects in their own, the teachers of secondary school tend to manifest multiple perspectives with similar levels concerning cognitive resources.