The original aims of our study were to investigate the dose-effect relationship of modafinil administration on working memory performance, in parallel with the measurement of plasma corticosterone in chronically-stressed mice, as compared to control mice. Memory performance was evaluated by spontaneous alternation in a T-maze. Vehicle or modafinil (8, 16 or 32 mg/kg) were administered after or without chronic stress (immobilization and exposure to light) for 15 min/day over a period of consecutive 14 days. Immediately after behavioral testing, blood was sampled to measure plasma corticosterone levels. Under non-stress conditions, corticosterone significantly increased with 16 and 32 mg/kg modafinil administration. Interestingly, optimal working memory performance was revealed at the 16 mg/kg dose. Moreover, no correlation was evidenced between working memory performance and plasma corticosterone level in modafinil-treated animals. Under stress conditions, corticosterone level was lowered at 8 mg/kg and remained unchanged at 16 and 32 mg/kg modafinil. An optimal working memory performance was evidenced at 8 mg/kg, which indicated a decrease in the efficiency threshold of modafinil under stress. Furthermore, an inverse correlation emerged between working memory performance and corticosterone level. Our study evidenced for the first time the interaction between stress and memory, in the emotional modulation of working memory performance, as a function of the administered dose of modafinil.