This paper addresses the issues of admissibility analysis and exponential passivity using sampled-data control for a class of singular networked cascade control systems (NCCSs) with time-varying delays and external disturbances. Based on the new augmented Lyapunov–Krasovskii functional which considers all the available information about the actual sampling pattern, a new set of delay-dependent condition is obtained to ensure the singular networked cascade control systems to be regular, impulse-free, stable, and exponentially passive. Based on the derived condition, the sampled-data control problem is solved and an explicit expression for the desired cascade controller is given. If the given linear matrix inequalities are feasible, then corresponding gain parameters of the designed cascade control will be determined. Finally, a numerical example based on a power plant boiler–turbine system is provided to demonstrate the effectiveness of the developed technique.