<p>In the nervous system, action potentials (APs) that propagate along axons are the primary carriers of encoded information. On the basis of the Hodgkin-Huxley model, this study constructs a model of a myelinated cortical axon to investigate the conduction dynamics of action potentials under sinusoidal and synaptic-like random current stimulation. The results demonstrated that under sinusoidal input, the stimulation frequency (<i>f</i>_in) and amplitude (<i>A</i>_in) jointly regulated the frequency-locked mode (<i>r</i> = <i>f</i>_out/<i>f</i>_in) at the proximal axon. AP transmission was modulated by the internodal conductance (<i>κ</i>) and feedback conduction current (<i>I</i><Stack> _{<i>i</i>,inter} ^← </Stack>). The feedback conduction current suppressed proximal depolarization, thereby reducing the frequency-locked ratio <i>r</i>. In contrast, increasing <i>κ</i> (&gt; 0.2mS/cm²) synchronized frequency-locked behaviors of distal nodes with that of the proximal node. Under synaptic-like stochastic input, high-frequency truncation (ISI &lt; 20 ms) at the proximal axon and AP loss during propagation caused a progressive decrease in information entropy (<i>H</i>) along the axon. However, the feedback conduction current attenuated proximal high-frequency truncation and achieved entropy conservation between input entropy and axonal conduction entropy (<i>H</i>_axon = <i>H</i>_in = 4.3 bit) at a specific parameter (<i>λ</i> = 22 ms), enhancing transmission fidelity. Moreover, under certain conditions, the temperature maximizes the locking frequency within 27.25°C–30.75°C while simultaneously intensifying high-frequency truncation. This work reveals that the feedback conduction current optimizes axonal information transmission efficiency through a dual mechanism: suppressing proximal firing and maintaining entropy conservation.</p>