In the field of fuzzy utility mining, the characteristics of transaction time have been a widely studied topic in data mining. However, using a single-conditional threshold for all items does not suffice to reflect the true properties of items. This paper, therefore, proposes a multi-conditional minimum threshold approach which considers the temporal behavior, the importance in items, and the usage of linguistic terms to mine high temporal fuzzy utility itemsets. The multi-conditional minimum thresholds are utilized to assist users in deciding appropriate standards for itemsets in mining when its contained items have different importance. An extended two-phase model and a corresponding mining approach are designed to handle the temporal fuzzy utility mining problem with the multi-conditional minimum thresholds. Finally, the results from the experimental evaluation show the effectiveness of the proposed approach under different settings.