<p>This paper investigates a distributed optimization problem for multi-agent systems with unknown cost functions and actuator saturation constraints. To address the integrator windup phenomenon induced by actuator saturation, a distributed extremum-seeking-based optimization algorithm incorporating anti-windup compensation mechanism is proposed. An event-triggered communication scheme is introduced in the designed algorithm for adaptive regulation of neighborhood information exchange frequency. Due to the multi-time-scale structure of the designed distributed algorithm, singular perturbation techniques are employed to achieve semi-globally practically asymptotic convergence of the decision variables of all agents to the optimal solution. Finally, simulations verify the effectiveness of the proposed algorithm.</p>