Purpose: To examine the reproducibility and sensitivity of a self-paced field running test (SFT) in monitoring positive and negative changes in endurance performance. Methods: A total of 27 (11 women) recreational runners participated in a 6-week training intervention. The intervention was divided into a 3-week baseline period, a 2-week overload period, and a 1-week recovery period. An incremental treadmill test was performed before the baseline period and a 3000-m running test before and after all periods (T1–T4). In addition, the participants performed once a week an SFT (SFT1–6), which consisted of submaximal (6 + 6 + 3-min test at perceived exertion of 9/20, 13/20, and 17/20) and maximal sections (6 × 3-min intervals at maximum sustainable effort). The associations between the incremental treadmill test and the SFT1 performance were examined with the Pearson correlation, and the intraclass correlation was analyzed for the parameters of SFT1 to SFT3 sessions during the baseline period. The repeated-measures correlation was calculated for the 3000-m speed at T1 to T4 and the corresponding speeds at SFT. Results: Significant associations (r = .68–.93; P < .001) were found between the speeds of the SFT and the peak and lactate threshold speeds of the incremental treadmill test. Intraclass correlations varied between .77 and .96, being the highest for the average speed of 6 × 3-minute intervals. Repeated-measures correlation was significant (P < .05) for the 9/20 (r = .24), 13/20 (r = .24), and 6 × 3-minute intervals (r = .29). Conclusions: The SFT seemed a reproducible method to estimate endurance performance in recreational runners. The sensitivity to track short-term and small-magnitude changes in performance seems more limited and might require more standardized conditions.