Banana production in Guadeloupe was developed on intensive monoculture design with high level of inputs. Currently, this production is cutting through an economic and an environmental crisis. With this crisis, solution possibility lies in proposing sustainable innovative cropping systems. Evaluation and design of innovative systems require the use of tools for modeling and assessment. The research unit ASTRO of the French National Institute for Agricultural Research has developed a program to design banana innovative cropping systems. First results modelized the existing farm diversity through a typological and systemic approach. Then a panel of experts designed a set of innovative cropping systems. Assessment is a key step in a prototyping framework (i) to provide action effects (ex ante assessment); (ii) to identify causes (diagnosis); (iii) to monitor a process; (iv) to take stock of actions (ex post assessment). These assessments are essential in order to make a decision (piloting, proposed actions, improvements) or to communicate. For our case, we chose a qualitative model to assess the overall sustainability of analternative cropping system: the model MASC (MultiAttribute Assessment of the Sustainability of Cropping systems) (Sadok et al., 2009). First, we adapted MASC to banana production in Guadeloupe, changing the decision tree structure (representing sustainability), the utility functions, some methods used to calculate the input attributes, and the threshold for qualitative transformation of inputs. The sustainability of 18 prototypes was assessed taking into account the six-class typology. The data needed came from interviews and results from the mechanistic model BANAD (Blazy et al., 2010). MASC allowed identifying different sustainable issues through each farm type. Among the 108 pairs "prototype*farm type" assessed, 51 scored a low level of sustainability, 54 a medium level, and three a high level of sustainability. One of the prototypes gets a high level of sustainability in a farm type whereas this same prototype scores a low level in another farm type. Also, two prototypes seem promising, scoring at least a medium level of sustainability among all the farm types. Our results show that considering the diversity of farms is important in the assessment process. Alternative cropping systems do not generally have the same effects through different farm types. Furthermore, the promising prototypes have to be tested in situ to assess their agronomic and functional performances. Stakeholders should take part in the adapting process of MASC. MASC was appropriate for its flexible structure suitable to be adapted to specific contexts. It can discriminate satisfactorily among a wide range of innovative cropping systems.