This paper aims to deal with existing uncertainties in different levels of information regarding coastal risk assessment from data acquisition and analysis, to the modeling and representation of risk zones. Risk assessment techniques require integrating several sources of data to provide a coherent and complete vision of potential risk regarding the phenomenon under study. This includes assessing possible damage on environmental, economical and social features as well as human-life losses. This fundamental information can then be analysed at a higher hierarchical level to choose appropriate actions and strategies to protect the region, its environment, the people and their assets in an optimal way. Complete and high quality data and information are mandatory in this regard to perform accurate assessment and efficient decision-making. Typically, data are likely to be collected and analysed by different authorities or organizations, with different levels of resolution and quality. Uncertainties exist and propagate from the collection, capture, storage, analysis and representation of spatial data to their interpretation and decision-making processes. Uncertainties can also appear as vagueness in boundary zone, ambiguities in linguistic terms, fuzziness in process interpretation, doubt in existence of a spatial object, or a combination of them. Today, ignoring uncertainty in data analysis and decision-making procedures is not considered as an efficient practice anymore. One dimension of uncertainty in coastal risk assessment is originating from risk zones representation. Traditionally, risk zones are represented by polygons that can be defined regarding to stakeholdersÂ’ interests or national census segments. Polygons are separated by well-defined boundaries while the degree of risk is attributed homogenously within each polygon considering multiple criteria. However, the way to define the shape of polygons differs among experts depending on their objectives. Likewise, the method to calculate and assign the degree of risk to each polygon is a challenging issue. Moreover, in reality, risk value changes continuously from one point to another. Thus, representing the transition from one zone to another zone with a crisply-defined boundary gives misleading insights of the risk degree of each region to decision-makers. Furthermore, risk has hierarchical characteristics due to the inherent needs and interests of different participants working for different organizational levels. For instance, their interests may be in an object such as a port, certain buildings, or more global like a census track, a city or even a state or country. In this regard, risk zones are complex objects with uncertain boundaries resulting from the fact that their definitions are vague and multi-scale. The flexibility of fuzzy set theory to express risk value, consistent with human reasoning, together with possibility of dealing with uncertainties suggests that it as an efficient solution for spatial representation and communication of the risk. This paper proposes an algorithmic approach based on fuzzy set theory to deal with the problem of ill-defined boundaries of risk zones. Then, a fuzzy object aggregation approach is proposed for multi-scale fuzzy representation of risk zones. Finally, the proposed approach is applied to coastal risk representation in Gaspe region, in Eastern Quebec, Canada for validation purpose.