Association rules mining is used to find interesting relations such as frequent patterns, correlations, and associations from data sets. Some association rule mining algorithms use data transactions in Boolean, categorical, or quantitative data. Meanwhile, a soft set theory is a mathematical tool often used to deal with data uncertainty. A soft set has excellent potential for application in many directions. The idea of mining association rules on transactional data based on a soft set was initiated by Herawan and Deris. After that, Feng Feng et al. provide further detailed insight into that idea. This idea provides new directions for implementing soft set theory in data mining areas. Whereas in the real world, many problems contain uncertainty, imprecise, and noise, represented by fuzzy soft sets. This paper proposed the approach of association rules mining based on a fuzzy soft set. For this purpose, new objective measures have been defined, such as fuzzy soft support ( fs-sup ) and fuzzy soft confidence ( fs-conf ). The experiment on breast cancer dataset showed that the objective measures properly facilitate the mining association rule of a fuzzy soft set.