Practice-based experiences documenting development and implementation of nutrition and health surveillance systems are needed. To describe processes, methods, and lessons learned from developing and implementing a population-based household nutrition and health surveillance system in Guatemala. The phases and methods for the design and implementation of the surveillance system are described. Efforts to institutionalize the system in government institutions are described, and illustrative examples describing different data uses, and lessons learned are provided. After initial assessments of data needs and consultations with officials in government institutions and partners in the country, a population-based nutrition surveillance system prototype with complex sampling was designed and tested in 5 Guatemalan Highland departments in 2011. After dissemination of the prototype, government and partners expanded the content, and multitopic nutrition and health surveillance cycles were collected in 2013, 2015, 2016, 2017/18, and 2018/19 providing nationally representative data for households, women of reproductive age (15–49 y), and children aged 0–59 mo. For each cycle, data were to be collected from 100 clusters, 30 households in each, and 1 woman and 1 child per household. Content covered ∼25 health and nutrition topics, including coverage of all large-scale nutrition-specific Interventions; the micronutrient content of fortifiable sugar, salt, and bread samples; anthropometry; and biomarkers to assess annually, or at least once, ∼25 indicators of micronutrient status and chronic disease. Data were collected by 3–5 highly trained field teams. The design was flexible and revised each cycle allowing potential changes to questionnaires, population groups, biomarkers, survey design, or other changes. Data were used to change national guidelines for vitamin A and B-12 interventions, among others, and evaluate interventions. Barriers included frequent changes of high-level government officials and heavy dependence on US funding. This system provides high-quality data, fills critical data gaps, and can serve as a useful model for others.