Most parents tend to attribute the poor performance of their children to the head teacher’s leadership style. The study sought to find out the relationship between democratic leadership style and students’ academic performance in Garissa township, Garissa County. This study was guided by Contingency and performance theory based on quantitative approach to collect data. The study adopted a correlational research design that sought to measure the relationship between the variables. The study targeted a population of 750 teachers and head teachers from 30 public secondary schools in Garissa Sub County. The sample was 200 participants consisting of 192 teachers and 8 head teachers. The sample was obtained through stratified sampling to establish the number of head teachers and teachers to be included in the study sample. Simple random sampling was used to select 8 head teachers from 30 schools and 192 teachers from 720 teachers. Questionnaires were used; one for teachers and another for head teachers to collect data. Pilot study was carried out in two neighbouring schools which were not included in the study. The data was analyzed quantitatively using both descriptive and inferential statistics. The study established that leadership styles of Headteachers will influence performance of teachers. The data further shows that many teachers in the current schools have not stay long also these may be to some extent contributing to in experience to handle learning hence low academic performance. The study recommended that head teachers ought to be given refresher courses on appropriate leadership styles to be applied in their institutions. Sensitization is essential on mutual participation by all to meet the organizational targets. More so headteachers need to be aware of participatory leadership.