Although taxes on land and property have many desirable attributes, the challenge of ensuring completeness of tax rolls and currency of valuations preclude their effective use to support urbanization and service provision in many developing countries. The example of Kigali shows how building footprints and heights generated from high-resolution satellite imagery, together with sales prices and routine statistical data, allow to assess and improve coverage and design of property tax systems. We show that only 40% of potential land lease fee revenue (of US$ 4.9 million) was collected and that moving to 1% value-based tax would increase revenue almost 10 times while being less regressive than the current system. While this could allow reducing the tax burden for low-income groups, exemptions should be applied with caution based on careful empirical analysis.