Evaluating computer science (CS) teacher competencies is crucial for advancing CS education. This study aims to develop and evaluate a self-efficacy assessment tool for CS teaching competence among 1412 CS teachers nationwide in Thailand. Using structural equation modeling, the validity of the tool was examined across four domains: Content Knowledge and Skills (CK), Instructional Design (ID), Classroom Management (CM), and Professional Development (PD). The results demonstrated a good model fit and validity (CFI = 0.95, TLI = 0.95, PNFI = 0.85, RMSEA = 0.08, SRMR = 0.02, GFI = 0.97). Significant differences in self-efficacy were observed across the domains with a medium effect size (X[superscript 2](3) = 1233.26, p < 0.001, w = 0.29), with the highest self-efficacy in PD, followed by CM, CK, and ID. Teachers in their 30 s showed greater confidence in CM, while those with 5-9 years of experience exhibited the highest overall self-efficacy. Secondary school teachers expressed more confidence in ID and PD than primary school teachers. Additionally, teachers in smaller schools, categorized by the number of students, reported significantly lower self-efficacy in CK (p = 0.01), ID (p < 0.01), and CM (p < 0.01) than those in larger schools. These findings highlight the need for specific and continuing PD programs to address differences in teacher competencies and improve the quality of CS education in Thailand.