• In this paper, the p-norm fixed-time synchronization of delayed ICVNNs is investigated for the first time, in which the settling time estimation in terms of p-norm is derived. Our results are more flexible and adjustable than these results [36–38]. • Compared with the works [3,13,36-38,42], what we get herein is fixed-time synchronization, and these results are also applied to image encryption. • Different from the real-valued INNs [5–7,19,21] and first-order CVNNs [13,39-42]. In this paper, the neural model is defined in the complex field and has an inertial term, which broadens the application range of INNs and makes up for some previous literature results. This paper researches the problem of p-norm fixed-time synchronization for a class of delayed inertial complex-valued neural networks (ICVNNs). By using reduced-order transformation and separating real and imaginary parts of complex-valued parameters , the second-order ICVNNs can be converted into the form of first-order real-valued differential equations. Then some new flexible and adjustable algebraic criteria to ensure the fixed-time synchronization of ICVNNs are established by means of the non-smooth Lyapunov function and inequality analytical techniques. Moreover, the settling time of fixed-time synchronization is theoretically estimated, which does not depend on the initial value of systems. Finally, simulation examples and applications are presented to illustrate the validity and availability of the obtained results. [ABSTRACT FROM AUTHOR]