Approximately 1 in 26 children worldwide die before the age of five, with more than 50% of these deaths occurring in sub-Saharan Africa. In Kenya, despite gains in reducing under-five mortality from 52 to 41.9 deaths per 1000 live births between 2015 and 2022, this figure remains high: about 64,500 children die annually, largely from preventable deaths. Additionally, widespread variations in the impact of interventions on risk factors and socio-structural inequities across regions indicate the need for additional efforts to address attributable causes of child deaths and improve intervention effectiveness. Siaya County is one of the leading regions in under-five mortality, with 67.4 deaths per 1,000 live births in Kenya. The current study determined the predictors of under-five mortality in Siaya County between 2015 and 2020. Specifically, it estimated the under-five mortality rate, and survival probability and examined the roles of maternal healthcare-seeking behavior and socio-demographic factors on under-five mortality trends in Siaya County. The study analyzed secondary data obtained from the Siaya Health Demographic Surveillance System from 2015 to 2020, a population-based longitudinal survey. The basic information on demography and health was collected semiannually using structured questionnaires from the population of interest. A total of 24452 under-five children were enrolled in the study, with a male-to-female ratio of 1:1. There were 1,540 (6.3%) deaths with a mortality rate of 62.98 deaths per 1,000 live births, of which 41.8% were neonatal deaths. The annualized rate of change of U5MR in the HDSS site was approximately 70^% with a survival probability mean age of 7 months. Cox proportional hazard was used to examine the effects of the predictors on under-five mortality at p-value < 0.05. In Cox multivariable analysis, the risk of death among females was 18 times (HR 0.82 CI=0.74, 0.91 p <0.001) lower than male children. Under-five mortality decreases with increasing educational attainment of the mother (HR 0.40 CI 0.23, 0.70 p <0.001) and number of ANC visits (HR 0.48 CI 0.26, 0.88 p= 0.018). Delivery in the hospital (p= 0.001) birth over age 18 years (p= 0.034) and use of a modern or pit latrine (p=0.03) improved survival. In conclusion, the mortality rate among U5s remains high and is mainly attributable to preventable causes. These observations reinforce the need to strengthen multiple components at the community and institutional levels. Interventions would target increasing uptake of preventive healthcare, risk reduction, and education attainment. The findings from this study are relevant to policymakers as they offer short-term solutions and integrated efforts to address the challenges.