This paper presents a reinforcement learning approach for optimized drone engine performance evaluation, focusing on turbojet engines with nonlinear system dynamics. Traditionally, classical PID controllers have been used, with optimal tuning derived from methods like the Riccati equation or Linear Quadratic Regulator (LQR). However, these methods lose effectiveness when the system deviates from linearization points and as physical characteristics change over time due to aging and wear, requiring manual retune and increasing maintenance costs. To overcome these limitations, we propose an adaptive PID controller optimize using the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm. Simulation results in Matlab Simulink demonstrate the proposed method’s ability to maintain optimal control performance under nonlinear and time-varying conditions, highlighting its potential for enhancing drone engine reliability and efficiency.