Purpose Using exercise protocols at a fixed rating of perceived effort (RPE) is a useful method for exploring the psychophysical influences on exercise performance. However, studies that have employed this protocol have arbitrarily selected RPE values without considering how these values correspond to exercise intensity thresholds and domains. Therefore, aligning RPE intensities with established physiological thresholds seems more appropriate, although the reliability of this method has not been assessed. Methods Eight recreationally active cyclists completed two identical ramped incremental trials on a cycle ergometer to identify gas exchange threshold (GET). A linear regression model plotted RPE responses during this test alongside gas parameters to establish an RPE corresponding to GET (RPEGET) and 15% above GET (RPE+15%GET). Participants then completed three trials at each intensity, in which performance, physiological, and psychological measures were averaged into 5-min time zone (TZ) intervals and 30-min ‘overall’ averages. Data were assessed for reliability using intraclass correlation coefficients (ICC) and accompanying standard error measurements (SEM), 95% confidence intervals, and coefficient of variations (CoV). Results All performance and gas parameters showed excellent levels of test–retest reliability (ICCs =  > .900) across both intensities. Performance, gas-related measures, and heart rate averaged over the entire 30-min exercise demonstrated good intra-individual reliability (CoV =  < 5%). Conclusion Recreationally active cyclists can reliably replicate fixed perceived effort exercise across multiple visits when RPE is aligned to physiological thresholds. Some evidence suggests that exercise at RPE+15%GET is more reliable than RPEGET.