The current study aimed to examine the effect of verbal instruction on explosive force production and between-session measurement reliability during maximal voluntary contractions of knee extensors. Following familiarization, 20 healthy males performed 3 maximal contractions with a “hard-and-fast” instruction and 3 maximal contractions with a “fast” instruction during 2 test-retest sessions. Knee extension maximal voluntary force (Fmax) and the maximal rate of force development (RFDmax) were measured. Maximal electromechanical delay (EMDmax), and the maximal rate of muscle activation (RMAmax) of quadriceps muscles were determined. No significant effect of instruction was observed on Fmax (p > 0.05). The RFDmax and RMAmax were significantly higher with the “fast” compared to the “hard-and-fast” instruction (36.07%, ES = 1.99 and 37.24%, ES = 0.92, respectively), whereas EMDmax was significantly lower with the “fast” instruction compared to the “hard-and-fast” instruction (-3.79%, ES = - 0.29). No significant differences between test and retest measurements were found (p < 0.05). However, the reliability of the RFDmax was higher with the fast instruction compared to the hard-and-fast instruction (CV: 7.3 vs. 16.2%; ICC: 0.84 vs. 0.56). Besides, the RFDmax was associated with the RMAmax and EMDmax with a significant effect of instruction. Data showed that the instruction given prior contracting muscle affected explosive force production and associated neuromuscular variables. As a result, the “fast” instruction may be preferred in the assessment of explosive force capacity of skeletal muscle during maximal efforts.