The event-driven model-free control is proposed in this article, which deals with the ‘trade-off’ between computational cost and system performance. Model-free controllers demand low computational resources and have high robustness, which is especially suitable for embedded systems with complex dynamics and/or affected by disturbances. The proposed method is implemented in two distinct nonlinear Multiple Input, Multiple Output (MIMO) motion models: a quadrotor model hovering in different weather circumstances and a vehicular friction control model operating in variable road surface conditions. Under the time- and event-driven schemes, the model-free control is compared with standard control strategies in various realistic scenarios.