Life Cycle Assessment (LCA) is a quantitative eco-design tool well recognized and useful for an assessment of environmental impacts of a product or service. However, an effective eco-design approach does not only require an assessment method but also a method which helps economic actors by providing guiding information to act towards environmental performances since they have generally multiple technical options to perform their activities. Sensitivity Analysis (SA) is an interesting method to compare these multiple options in a systematic manner and which can help in understanding the influence of input parameters on the output. SA is already used in LCA studies, as recommended by the ISO 14040 standard, by choosing options to vary one by one, with the aim of testing robustness of results, by varying parameters one by one after LCA has been conducted. However, this type of SA has two weaknesses. The first is that varied parameters are arbitrary chosen, generally according the expectation that it can be influent. This expectation can be based on experience but is however very subjective, and can conduct to neglect influent parameters that were not expected so. The second weakness is that varying parameters one by one is assuming that all parameters are linear and independent, without checking that there are, and thus taking the risk to be blind about important interaction effects. In the purpose of constructing an action-oriented eco-design approach, it is proposed to combine SA and LCA tools during the first step of the LCA method, i.e. the system modeling step. The purpose is to rank and quantify the influences of the each input parameters (or group of parameters), without neglecting one a priori. These parameters are grouped in main categories reflecting the action possibilities of the economic actor: technological parameters, i.e. those that economic actors can directly control; environmental parameters which cannot be directly controlled and depend on local conditions; and methodological parameters that are not at all controlled by the economic actor. The LCA and SA methods are combined as follows: 1) choice of the models, 2) identification of the input parameters, 3) Application of the Morris method to identify and rank influential parameters, to determine the tendency (increasing or decreasing), and to simplify calculations by eliminating non influent parameters from further analysis 4) Application of the Sobol method to quantify the influence of parameters on the LCA impact indicators. This LCA and SA combination is applied to hemp based insulation products for building construction. The considered economic actor is the project manager of the building. The present work benefits from results from a previous study, for which the economic actor was the farmer. Thus only parameters founded influent for the farming system, have been kept for the present study. The system is a cradle to gate system: it includes agricultural cultivation, primary and secondary transformations of the hemp to various possible insulation products. Varied parameters (their domain and type of variation) are described all along the supply chain. Ecoinvent life cycle inventory database were used for all background processes of the studied system and the production of all inputs. To convert inventory flows to impact indicators, the CML 2001 characterization method and the cumulative energy demand were used. Results show that technological and methodological parameters are the most influent on the variance of the studied impact categories, meaning that the choices of the economic actors have influence on global performance of the system. Parameters from the secondary transformation sub-system are found to be the most influential, except for eutrophication and ozone depletion for which the agricultural sub-system is found to be the most influential. Influences are also quantified by individual parameters as well as their interactions. LCA results are given distinctively by most or less favorable choices made by economic actor.