ABSTRACT: Since the mid 20th century, one of the best horizons for stratigraphic correlations in clastic and carbonate systems has been what is now commonly identified as the “maximum flooding surface.†However, various definitions for the term “maximum flooding surface†have appeared over the years, calling on a range of sedimentary processes to form the surface depending on the depositional system. Disagreement over the definition of the maximum flooding surface is largely due to ambiguous characteristics, assorted identification techniques and conflicting genetic interpretations. Multi-valued logic (or fuzzy logic) methods can be used to address some of these problems of uncertainty. This paper discusses how the multi-valued logic methods of “fuzzy classifier design" and “fuzzy inference systems" are used to manage subjectivity in carbonate sequence stratigraphic models. Multi-valued logic methods are used to 1) quantify sedimentary characteristics through fuzzy classifier design to elucidate subjective identification and to facilitate repeated recognition of the maximum flooding surface in rock descriptions and to 2) develop a fuzzy inference system to model the development of the surface. The fuzzy inference system argues that in carbonate systems the time of maximum sediment starvation and maximum water depth are discrete events. The surface of maximum starvation is recorded as a nearly synchronous horizon associated with the maximum rate of sea level rise, whereas the maximum bathymetric event is highly diachronous due to the interplay of rates of sediment accumulation and accommodation space.