This study aims at evaluating the neuromuscular adaptations occurring during submaximal prolonged cycling tasks. In particular, we want to assess changes in surface electromyographic (sEMG) signal recorded during a pedaling task, performed by six subjects on a cycle-simulator at a constant power output, until voluntary exhaustion. Task failure was defined as the instant the subject was no longer able to maintain the required task. Electromyographic activity was recorded from eight muscles of the dominant leg and burst characteristics of sEMG signals were analyzed in order to assess the changes in muscle activity level produced by the occurrence of neuromuscular fatigue. In particular, three features were extracted from the sEMG signal for each burst: amplitude, location of the maxima and mean profile of the burst envelope. We have reported an increase in the amplitude parameter for all subjects only for Vastii while bi-articular muscles presented a high variability among subjects. Also the location of the maximal values of the mean envelope of the bursts was found to change when considering bi-articular or mono-articular muscles. The envelope profile was found not to be subject to alterations when comparing the end of the task with the beginning. We speculated that neuromuscular fatigue induces changes essentially in the mono-articular muscles which produce power. This phenomenon is highly correlated with the adopted pedaling strategy which, being not constrained, induces subjects to express the maximal power in the downstroke phase, related to knee extension and involving mainly mono-articular muscles.