Teachers' self-efficacy is belief in their capacity to teach successfully. Globally, studies indicated the importance of teachers' self-efficacy, while studies conducted in Africa suggested that there could be differences in teachers' self• efficacies across demographic variables. Low teachers' self-efficacy in Kisumu County was suggested by studies conducted in parts of Kisumu County showed and by the high rate of transfer requests and resignations. However, Kisumu County Education and Teachers' Service Commission offices had no published seminars or workshops conducted to teachers to address teachers' self-efficacy. The purpose of this study, therefore, was to investigate the influence of selected demographic variables on teachers' self-efficacy in public secondary schools of Kisurnu County, Kenya. The study objectives were: to establish the influence of gender on teachers' self-efficacy; determine the influence of marital status on teachers' self-efficacy; find out the influence of school category on teachers' self-efficacy; establish the influence of age on teachers' self-efficacy and explore the influence of teaching experience on teachers' self-efficacy. Bandura's Social Cognitive Theory guided the study; and the study adopted Mixed Method Approach and within it, Concurrent Triangulation Design was used. Study population was 1790 teachers in 143 public secondary schools. The quantitative sample was 327 teachers selected using stratified random sampling based on Israel's formula, while the qualitative sample was 13 teachers selected using strategic sampling. Interview schedule and questionnaire were used to collect qualitative and quantitative data respectively. The pilot study used 2 and 33 respondents for interview schedule and questionnaire respectively drawn from 15 schools, which were not sampled in the study survey. Two experts in Educational Psychology ensured face validity by examining the data collection instruments. Construct validity was ensured by computing convergent (r = 0.564) and discriminant validity (r = -0.325). Internal reliability for the questionnaire was computed and found to be Cronbach's a = 0.9976 while trustworthiness of qualitative data was ensured through triangulation, iterative questioning and fillmg the sampling frame. Qualitative data was analyzed using Thematic ysis while quantitative data was analyzed using descriptive and inferential Sl:!::!..<t!Cs such as Multivariate Analysis of Variance. The quantitative findings ed that only school category had a significant influence on teachers' self• effi Wilk's 1'. (2, 324) = 0.893, p = 0.000) and all its domains (student en= gement. F (2, 324) = 11.498, p = .000; instructional strategy, F (2, 324) = 8A3_. p - .000: and classroom management, F (2, 324) - 10.173, p - .000). Qualitative data revealed that gender, marital status, school category, age and experience influenced teachers' self-efficacy. ln addition, qualitative data revealed that the influence of age on teachers' self-efficacy increased up to between 40 and 50 years old and then declined. The study concluded that self• efficacy is highest among male, married, national school and experienced teachers aged between 40 and 50 years. The study recommends to the Ministry of Education, Science and Technology that Form One selection should be random and not based on marks at Kenya Certificate of Primary Education.