This paper uses research fromneuroscience and the psychology of mathematics to arrive atuseful recommendations for teaching mathematics at primarylevel to poor students in developing countries. Theenrollment rates of the poorer students have improvedtremendously in the last decade. And the global NetEnrollment Ratio (NER) has improved since 2001 from 83.2percent to 90-95 percent except in Sub-Saharan Africa andSouth Asia. Making teaching of math and other subjectsefficient for the poor in developing countries is a greatchallenge, particularly in south Asia and Sub-SaharanAfrica. Many developing countries have explored new means ofteaching math and other subjects. Mongolia changed itsmathematics education, aiming to build a new set ofpriorities and practices, given the abandonment of earliertraditions. Similar to international trends of the time,South Africa in the 1990s extensively applied theconstructivist learning philosophy which relied onexploration and discovery, with little emphasis onmemorization, drill, In conformity with a belief thatteachers could develop their own learning programs, therewas virtual absence of a national or provincial syllabus ortextbooks. Students were expected to develop their ownmethods for arithmetic operations, but most found itimpossible to progress on their own from counting to actualcalculating. This study integrates pertinent research fromneuroscience and the psychology of mathematics to arrive atrecommendations for curricular and efficient means ofmathematics instruction particularly for developingcountries and poor students at primary level. Specifically,the latest research in neuroscience, cognitive science, anddiscussions of national benchmarks for primary schoolmathematics learning, form the basis of our recommendations.These recommendations have a reasonable chance of working inthe situational contexts of developing countries, with theirtraditions and resources.