Fatigue crack propagation is a stochastic phenomenon due to the inherent uncertainties originating from material properties, environmental conditions and loads. Stochastic processes offer an appropriate framework for modelling crack propagation. Indeed, these processes enable us to include certain variabilities. In this work, we propose to model the crack propagation mechanism with piecewise-deterministic Markov processes and typical deterministic crack laws. Conventional equations proposed in the literature seem inadequate for describing the entire fatigue crack trajectory, especially when the crack extends in a rapid manner. We propose a regime-switching model to overcome this challenge, in which the propagation is randomly divided into two parts. Each of these parts is governed by a deterministic equation whose parameters are randomly selected in a finite state space. We adjust the parameters from real data available in the literature. The behaviour of the propagation is well captured from the propagation phase until the ''fast crack propagation" phase leading to failure. Thus, the proposed switching model enables us to understand the change between these two phases of propagation. Statistical observations and numerical simulations demonstrate the efficiency of our approach to model fatigue crack growth.