In this paper, the interbasin motion (IBM) approach is applied to studying dynamics of conformationally constrained peptides, being extended to a nonideal contact of the system with a thermal bath. The coupling of the system with the thermal bath is expressed in terms of a memory function. The aim of the present study is twofold. First, we present a dynamical diagnosis of the three hexapeptide variants with the main focus on the transitions between basins rather than between individual states. Second, the present study is intended to pinpoint a way for extracting useful information about the strength of the system-solvent coupling and how this interaction affects the propensity of relaxation towards the native state. We show that a slight variation of the value of the memory friction parameter may induce a sizable modification of the relaxation time. In addition, the change of the memory friction parameter produces alterations on short time scales among the population distributions. Especially, high energy basins seem affected the most. In the Markovian limit, the basin populations computed within the IBM model are compared to those obtained by using state-to-state transition rates in the full master equation approach. The two methods yield similar results when the separation of time scales between intra- and interbasin dynamics is completely achieved. (C) 2003 American Institute of Physics.