Further agricultural intensification is expected to meet the on-going demographic transition and changes in consumption habits. There is therefore a need to identify contexts, options and conditions for intensive but sustainable land uses to develop. This question is of particular importance under contexts of smallholder farming and pressure on natural resources, such as in Asian humid tropics, where conservation agriculture (CA) is considered by several authors a leeway for sustainable intensification of agriculture. Within this context, our work aimed at preliminary assess to which extent CA fits the needs and constraints of smallholder farmers in a mountainous region where the high level of pressure on natural resources adds-up to a high poverty rate. We firstly implemented a diagnosis of sustainability gaps associated with the regional diversity of agricultural systems. To do so, we applied an original methodological approach combining diverse classification tools to maize based agricultural systems on sloping land. We identified 5 farm types and 7 maize cropping systems types, strongly contrasted in terms of management practices, performances and sustainability issues. The local diversity in maize cropping systems resulted from multi-scale interactions between territory-related factors, farm/farmer characteristics and field biophysical conditions. Within the context of our study, territory-related factors were of prevailing importance over plot biophysical conditions and farm/farmer related characteristics to drive diversity of cropping systems. Increasing economic profitability and agronomic productivity were compatible goals. However, tradeoffs were detected between economic profitability and mining of soil fertility. We subsequently investigated to which extent data collected at demonstration sites under project management can be used to determine whether CA has chance to become more productive and profitable over a 2-year period. Within the context of our study, CA associated with sufficient fertilization levels did not depress yields the first year after conversion. The second year, CA significantly increased maize productivity and agronomic efficiency. Such improvement of agronomic performances was not sufficient to ensure better economic performances than conventional agriculture over two years, due to the increase in production costs associated with transition to CA. Beyond those specific results, our work questions how CA-demonstrations setup by development project can provide valuable contribution to (i) the analysis of process driving agronomic and economic performances of demonstrated cropping systems and (ii) the investigation of cropping system agronomic and economic performances across variations in biophysical and economic conditions. Operational constraints and underlying demonstration goals limit possibilities to improve demonstration design. Coupling demonstration with modeling therefore appears the most effective way to improve scientific output from such demonstrations. Bio-economic farm model would therefore allow refining ex-ante investigation of CA economic attractiveness for smallholder farmers considering diverse decision timelines and differentiated dynamics of cropping system performances over time. Other scientific perspectives include the assessment of inter-annual variability in cropping system performances, including environmental outputs and the investigation of capacity of conservation agriculture to buffer this variability. Last, our results have practical implications for sustainable agricultural intensification in tropical humid highlands of Vietnam. Typologies of agricultural systems and field-scale analysis of cropping system performances allow building new assumptions for further cropping system prototyping and farm systems to be targeted in CA extension. To make conversion to CA worth for farmers within a period of two years, a major assumption is the requirement for cover plant valorisation, implementation of stepwise transition to CA, and/or tailoring incentives to support smallholder farmers in overcoming costs associated with conversion to CA.
