According to the World Health Organization (WHO), the high number of maternal deaths in some areas of the world reflects subnational inequities in access to maternity health. In Siaya County, Kenya, an estimated 34% of pregnant women still delivered at home in the year 2018. Yet it is still unclear whether the problem is the lack of capacity in the available maternity services or if within the predominantly rural county, choice of transportation modes (walking versus motorized), combined with the limited road networks, poor road quality, and diverse terrain is hindering pregnant women’s physical access to the existing maternity health services. Furthermore, it is currently unknown whether being outside the geographic catchment area of existing maternity health services, influences pregnant women’s health-seeking behaviour. As such, the general objective of the current study was to model the impact of geographic inequalities on the utilization of maternity health services in Siaya County. The present cross-sectional study was a health system bottleneck analysis that used publicly available geospatial data, including the catchment area population reported by each facility and the total number of health facility deliveries per administrative ward in 2018 from the web-based District Health Information System (DHIS2). In this study, two independent variables, accessibility coverage, that is, the number of pregnant women who could access maternity services within 1-hour of travel time per ward, and geographic coverage, that is, the number of pregnant women within a 1hour geographic catchment of maternity health facilities per ward were simulated using a least-cost path algorithm in line with Tanahashi's framework of evaluating health service coverage. AccessMod (version 5.0), which is a free and open-source toolbox developed by WHO, in combination with a third-party geographic information system (GIS) software, ArcGIS (version 10.5), were used for geospatial modeling. Moreover, using a free software environment for statistical computing and graphics, R (version 3.5.3), a Zero-inflated Poisson regression model sufficed to test the relationships between the outcome variable (total number of health facility deliveries per administrative ward) and the aforementioned independent variables (accessibility and geographic coverage) at 95% confidence level. For all the study variables, the common denominator was the total number of pregnant women per administrative ward. More importantly, the current geographic inequalities evaluation revealed that approximately 70% of pregnant women across Siaya County could access existing maternity health services within an hour of traveling time. Overall, 54% of health facilities were estimated to be working above their capacity resulting in 37% of pregnant women being outside the geographic catchment area of the existing maternity health services network in Siaya County. Regardless, a p<0.05 suggests that being able to access a higher tier facility offering maternity health services by foot (CI0.339-0.347), was a better predictor of Skilled Birth attendance (SBA) in Siaya County. As such, the County Health Management Team (CHMT) needs to upgrade the quality of lower-tier maternity health services such as dispensaries, as pregnant women may value the quality of services regardless of the distance. Additionally, the available health development funding should be targeted toward the upgrade of hospital facility types working above their capacity, otherwise, the County government needs to also explore the option of constructing at least 32 tier two facilities in Siaya County, towards universal coverage. Future research should incorporate actual costs including the maternal lives saved or deaths averted into the model.