This paper deals with the adaptive reliable $\boldsymbol {H_{\infty }}$ optimization control problem for linear time-delayed systems (LTDSs) against time-varying actuator faults. A novel delay-dependent adaptive $\boldsymbol {H_{\infty }}$ performance index is described for LTDSs. An exponential function term has been successfully introduced to establish a novel actuator fault parameters estimation scheme. Moreover, the designed indirect adaptive controller includes both offline and online gains, which can be obtained by robust optimization technique and adaptive fault parameters estimation, respectively. By employing an improved Lyapunov–Krasovskii functional and applying some new-developed integral inequalities technique, some delay-dependent sufficient conditions with less conservatism are established to make the closed-loop systems exponentially stable and have the optimized adaptive delay-dependent $\boldsymbol {H_{\infty }}$ performance index. Finally, the effectiveness and superiority of the presented scheme are demonstrated by two examples.