This paper surveys natural computing models and methods with application to risk management in complex systems. The equivalence of risk minimization and effective control problems is shown. The general risk management problem is stated for complex systems under uncertainty. The structure of fundamental and applied risk management problems is described. The well-known natural computing methods are briefly considered with application to risk management by the criteria of formalism, universality, and learning capability. The scientific community’s preferences in natural computing models and methods for solving different classes of risk management problems are analyzed. Some promising approaches are outlined, which are currently underinvestigated according to the authors’ opinion.