Maize is a staple crop and underpins food security for Zambia. Maize productivity in Zambia is almost half the potential due to low uptake of conservation agricultural practices. This study tests the hypothesis of the trade-offs between agriculture productivity and long-term ecosystem services (ES) provision among maize farmers practising conventional agriculture on the one hand, and those practising conservation agriculture on the other hand, in Kafue district, Zambia. In addition, challenges which affect conservation agriculture uptake are assessed. Besides these challenges, the study notes that adoption of a new technology is also influenced by its efficiency, and therefore technical efficiency scores were estimated using the Stochastic Frontier Approach (SFA) to compare efficiency levels of the two agricultural systems. Kafue was purposefully selected as it is among the first districts where conservation agriculture was introduced. Through purposive and random sampling, the households surveyed were split into two distinct groups namely conservation agriculture (CA) farmers (treatment group) and conventional agriculture (CV) farmers (control group). The analysis significantly shows that farmers practicing CA have more knowledge than CV farmers about the capacity of conservation agriculture to reduce soil erosion, increase soil fertility, retain nutrients, mitigate pests and weeds and increase crop yield. However, both farming groups knew that CA helps conserve soils and that soil maintenance is important for food production. On the other hand, significantly, CV farmers knew more than CA farmers that CV reduces crop yield and increases soil erosion. On the other hand, CV farmers expressed a higher level of willingness to adopt CA practices than CA farmers who are unwilling to expand their area under CA. The study further shows that at least 55% of farmers practicing CA find inadequate labour to be the main challenge faced in CA. It appears that a policy that improves the farmers' knowledge on CA would help improve the uptake of CA. CA farmers were significantly more knowledgeable than CV farmers about the detrimental effects of CV, such as increasing air and water pollution, thus raising the need for training among CV farmers with emphasis on the effects of their farming system on the environment which affect the supply of ecosystem services. Efficiency is also a means of improving productivity hence the Stochastic Frontier Analysis (SFA) was employed to estimate technical efficiency levels in maize production. Using SFA, the study found that the technical efficiency of maize among CV farmers is 71.3% on average while that of CA farmers is 57.9% on average. Moreover, the study found that there was a significant difference (t=3.9854, P=0.0002) in the technical efficiency scores of the both CV and CA farmers. Nevertheless, the study also found that 77% of output variation among CA farmers can be explained by variation in technical efficiency. However, 33.4 % of total maize output can be explained by variation in technical efficiency among CV farmers. This means that CA farmers have a higher potential to increase their current output than CV farmers. Finally there was no significant difference in fertiliser usage between CA and CV farmers (t=1.3825, P=0.1700). Further, from SFA fertiliser responsiveness to maize output showed that a 1% increase in the use of synthetic fertiliser leads to 0.678% and 0% increase in maize output per hectare for CA and CV respectively at both 1% and 10% level of significance at the expense of water and air quality. Therefore, it can be concluded that there is a higher trade-off between maize production and water and air quality under conventional than under conservation agriculture.