Fiber-reinforced polymer matrix composites have excellent in-plane stiffness and strength properties, and are therefore ideal for usage in panels of aircraft wings or fuselage as well as launch vehicle case segments. Those thin plates or shell structures are often stiffened with many locally increased thickness regions, or beams of various cross-sectional shapes such as flat or T-shaped. Small defects in any of those stiffened regions would greatly reduce the structural performance as a whole. Locating such defects is time consuming because of the large extend of the panels as well as the number of stiffeners. Guided ultrasonic wave-based techniques could be applied for damage detection in large areas. However, the scattering characteristics of stiffeners are complex. In particular, when multiple stiffeners are present, the incident Lamb wave signal is altered with the passing of each stiffener. Thus, the goal of this work is to efficiently model Lamb wave propagation when multiple stiffeners are present, with and without defects, in an effort to identify useful signal features for damage detection. To this end, the so-called global-local method is used for Lamb wave modeling. The global functions are used to represent the nominal composite region – parameters are obtained by means of waveguide finite element (WFE) method – and the stiffened region is represented by finite element discretization. With a recently developed coupling technique, a source problem, representing a surface-mounted transducer is coupled with multiple stiffener-scattering models to examine the transmission characteristics. The global-local model is validated by laboratory waveform measurements on a stiffened composite plate. The results from global-local method can then be used to efficiently determine the maximum number of stiffeners before the transmitted Lamb waves become too weak to identify defects.