The author obtains two solutions for the uncertainty problem in a multistep decision-making problem for a wide class of preference choice rules in a decision-making system. They are based on the principles of guaranteed and best results, respectively, with the criteria in the form of preferences on decisions defined by an explicitly specified utility function, which parametrically depends on a convex statistical regularity on the set of states and on the utility function on the consequences, which is determined to within a positive linear transformation.In considering a decision-making problem (decision problem (DP) in brief) in a decision-making system, which is a pair of a decision-maker (DM) and a situation of decision-making (SDM) (see [1]), the question arises about the criterion according to which the DM implements the procedure of choosing this solution from a set of possible ones. This generates the problem of uncertainty of decision procedure.Let us consider an approach to the problem of uncertainty for problems of multiple decision-making presented in the form that expands the set of random consequences (the so-called neo-Bayesian form [2]) to random (in a broad sense) consequences.A decision-making problem where the consequences are determined by an action (decision) and the state of nature (unobservable parameter) is said to be a parametric DP. Let us formalize it as follows.Definition 1. A scheme of a situation of decision problem (SSDP) is an ordered quadruple ( , , , ) X U g Q , where g for arbitrary nonempty subsets X U , , Q and is a mapping of Q´U onto X .