Abrupt transitions in ecosystems can be interconnected, raising challenges for science and management in identifying sufficient interventions to prevent them or recover from undesirable shifts. Here we use principles of network controllability to explore how difficult it is to manage coupled regime shifts. We find that coupled regime shifts are easier to manage when they share drivers, but can become harder to manage if new feedbacks are formed when coupled. Simulation experiments showed that both network structure and coupling strength matter in our ability to manage interconnected systems. This theoretical observation calls for an empirical assessment of cascading regime shifts in ecosystems and warns about our limited ability to control cascading effects.