A sporadic rule is a rare association rule (which has low support but high confidence). It is divided into two types: perfectly and imperfectly sporadic rules. The problem of mining perfectly sporadic rules has been completely solved on both transactional and quantitative database. However, some approaches finding imperfectly sporadic rules on the transactional database have just been showed. The problem of mining imperfectly sporadic rules on quantitative database has not been solved yet. Thus, the paper aims to find an absolute answer to the question by proposing a problem of mining fuzzy imperfectly sporadic rules with two thresholds and developing a MFISI (mining fuzzy imperfectly sporadic itemsets) algorithm to find fuzzy imperfectly sporadic itemsets with two thresholds. The development of MFISI algorithm is derived from MCISI algorithm for which mining imperfectly sporadic itemsets with two thresholds on transactional database.