Finite-time stability of a class of fractional-order complex-valued memristor-based neural networks with both leakage and time-varying delays is investigated in this paper. By employing the set-valued map and differential inclusions, the solutions of memristor-based systems are intended in Filippov’s sense. Via using Hölder inequality, Gronwall–Bellman inequality and inequality scaling skills, sufficient conditions to guarantee the stability of the system are derived when 0<α<12 and 12≤α≤1, respectively. Finally, two numerical examples are designed to illustrate the validity and feasibility of the obtained results.