Cognitive Fatigue (CF) is an important confound impacting cognitive performance. How CF is triggered and what are the features that make a cognitive effort perceived as exhausting remain unclear. In the theoretical framework of the Time-based Resource-sharing (TBRS) model (Barrouillet et al., 2004), we hypothesized that CF is an outcome of increased cognitive load due to constrained time to process ongoing cognitive demands. We tested this cognitive load-related CF hypothesis across 2 experiments manipulating both task complexity and cognitive load induced by the processing time interval. To do so, we used the TloadDback paradigm, a working memory dual task in which high and low cognitive load levels can be individually adjusted. In Experiment 1, participants were administered a high cognitive load (HCL, short processing time interval) and a low cognitive load (LCL, large processing time interval) conditions while complexity of the task was kept constant (1-back dual task). In Experiment 2, two tasks featuring different levels of complexity were both administered at the individual's maximal processing speed capacity for each task (i.e., short processing time interval). Results disclosed higher CF in the HCL than in the LCL condition in Experiment 1. On the contrary, in Experiment 2 similar levels of CF were obtained for different levels of task complexity when processing time interval was individually adjusted to induce a HCL condition. Altogether, our results indicate that processing time-related cognitive load eventually leads to the subjective feeling of CF, and to a decrease in alertness. In this framework, we propose that the development of CF can be envisioned as the result of sustained cognitive demands irrespective of task complexity.