Mulugetta Mekuria is a senior scien-tist with the Socioeconomics Program, based at the International Maize and Wheat Improve-ment Centre (CIMMYT) Southern Africa Regional Office based in Harare, Zimbabwe. Christian Thierfelder is a conservation agriculture specialist with CIMMYT Southern Africa Regional Office based in Harare, Zimbabwe. Abstract: Understanding factors affecting farmers' adoption of improved technologies is crit-ical to success of conservation agriculture (CA) program implementation. This study, which explored the factors that determine adoption and extent of farmers' use of the three principles of CA (i.e., minimum soil disturbance, permanent soil cover with crop residues, and crop rotations), was conducted in 10 target communities in 8 extension planning areas in Malawi. The primary data was collected using structured questionnaires administered to individual households. Triangulation with key informant interviews, field observations, and interactive discussions with farmers and farmer groups provided information behind contextual issues underpinning the statistical inferences. From a total of 15,854 households in the study areas, it is estimated that 18% of the smallholder farmers had adopted CA, representing an area of about 678 ha (1,675 ac; 2.1% of all cultivated land). Land area under CA constituted about 30% of total cultivated land among adopters. A random sample of 151 adopters and 149 nonadopters proportional with respect to adoption rates was drawn from various communi-ties and interviewed using structured questionnaires. A total of 30 key informant interviews were conducted with stakeholders including staff of Total Land Care, government extension workers, agroinput suppliers, and lead farmers. The first stage of the Heckman model showed that hired labor, area of land cultivated, membership to farmer group, and district influenced farmers' decisions to adopt CA. The second stage of Heckman model results suggested that total cultivated land, duration of practicing CA, and district influenced farmers' decisions to extend their land to CA. Our study can be used to show the agency and social structures that are likely to influence adoption and extent of CA. Future policy should address ways to provide access to information and long-term support to farmers to enable them to embrace the technology fully.