Background: Despite scaling up Isoniazid preventive therapy implementation in areas with a high prevalence of HIV and latent tuberculosis infection >30%, there is a paucity of data assessing adherence which is pivotal to END TB control and elimination. We sought to determine the adherence level and its correlates among people living with HIV initiated on IPT in selected hospitals in Kisumu Central, Kisumu County. Methods: A facility-based cross-sectional study was conducted at Jaramogi Oginga Teaching and Referral Hospital, Kisumu County Hospital and Lumumba Sub-County Hospital between June and July 2018. A random sample of PLHIVs aged ≥18 years, initiated on IPT between 2016 and 2018 were interviewed. Self-reported method was used to ascertain adherence. Data was collected using Commcare and analysed with STATA 14.0. A generalized linear regression model was used to generate the adjusted prevalence ratios and 95% confidence intervals. Results: Of 462 respondents, 282(61%) were females. The mean age of respondents was 37.9 [±10.4, SD]. Forty percent (40% [n=185; 95 C.I = (35.6%-44.6%) adhered to treatment. Respondents who had knowledge of latent tuberculosis infection were more likely to adhere compared to those who had no knowledge [aPR=1.6; 95%CI= (1.16-2.2), P=0.004]. Respondents who experienced IPT stock-outs were less likely to adhere as compared to those who experienced no stock-outs [aPR=0.15; 95%CI= (0.02-0.93); P=0.042]. Conclusion: The overall adherence level is sub-optimal against a set threshold (≥ 80%). Knowledge and IPT stock-outs were associated with adherence. Sustained awareness campaigns and uninterrupted supply of IPT would optimize on adherence.