This study compared the effects of linear (LP) and nonlinear (NLP) training periodization on cognitive functions, neurotrophic biomarkers [plasma brain-derived neurotrophic factor (BDNF), insulin-like growth factor-1 (IGF-1)], and cathepsin-B in patients with coronary artery disease (CAD). Forty-four patients with CAD reported to our laboratory on two occasions to undergo testing procedures before and after training sessions, and were then blindly randomized to NLP or LP for 36 training sessions. Visit 1 included blood samples and a maximal cardiopulmonary exercise testing to get maximal oxygen uptake (V̇o2peak). Visit 2 included cognitive functions assessment. Thirty-nine patients completed the study (LP: n = 20, NLP: n = 19), with no observed changes in cognitive performances after the training intervention in either group. IGF-1 concentration decreased in both groups (time-effect: P < 0.001), whereas BDNF concentration increased (time-effect: P < 0.05) without group interaction, and cathepsin-B did not change after the intervention. Associations were found between ΔV̇o2peak and ΔBDNF (R2 = 0.18, P = 0.04), and ΔIGF-1 and Δshort-term/working memory (R2 = 0.17, P = 0.01) in the pooled sample, with ΔIGF-1 and ΔBDNF accounting for 10% of the variance in Δshort-term/working memory. In the LP group, associations were found between ΔV̇o2peak and ΔBDNF (R2 = 0.45, P = 0.02), ΔBDNF and Δshort-term/working memory (R2 = 0.62, P = 0.004), ΔIGF-1 and Δshort-term/working memory (R2 = 0.31, P = 0.01), and ΔIGF-1 and Δexecutive function (R2 = 0.22, P = 0.04). This study indicates that linear and nonlinear training periodization led to an increase in BDNF, and a decrease in IGF-1, without change in cognitive function in individuals with stable CAD.NEW & NOTEWORTHY We used a novel and supervised iso-energetic training, integrating both moderate- and high-intensity aerobic exercises. Our findings indicate that greater variation in training load did not yield cognitive enhancements, although both protocols exhibited positive effects on brain-derived neurotrophic factor (BDNF) levels. Moreover, this study establishes a clear positive association between short-term and working memory and neurotrophic biomarkers. In addition, the independent predictive value of change in insulin-like growth factor-1 (IGF-1) on improvement in short-term and working memory highlight the close relationship between neurotrophic markers and cognition. Consequently, our results advocate for exercise training interventions targeting neurotrophic biomarkers to enhance cognitive function among individuals with coronary artery disease.