||Remoteness has a crucial role in risk assessments of megaprojects, resilience assessments of communities and infrastructure, and a wide range of public policymaking. The existing measures of remoteness require an extensive amount of population census and of road and infrastructure network data, and often are limited to narrow scopes. This paper presents a methodology to quantify remoteness using nighttime satellite imagery. The light clusters of nighttime satellite imagery are direct yet unintended consequences of human settled populations and urbanization; therefore, the absence of illuminated clusters is considered as evidence of remoteness. The proposed nighttime remoteness index (NIRI) conceptualizes the remoteness based on the distribution of nighttime lights within radii of up to 1,000 km. A predictive model was created using machine learning techniques such as multivariate adaptive regression splines and support vector machines regressions to establish a reliable and accurate link between nighttime lights and the Accessibility/Remoteness Index of Australia (ARIA). The model was used to establish NIRI for the United States and Canada, and in different years. The index was compared with the Canadian remoteness indexes published by Statistics Canada.