Active infrared thermography has gained increasing popularity for nondestructive testing and evaluation in various industrial fields, especially for composite structures. In this regard, thermal wave radar (TWR) imaging is recognized as the next-generation active thermography technology to obtain great resolution and depth range over the inspected objects. A critical aspect concerns the optimal test parameter selection to guarantee reliable quality assurance required for industrial products. In this work, single- and multiple-frequency TWR was investigated in a quantitative manner with the goal of optimizing the detection parameters in terms of probing range and lateral and depth resolution. The effects of test parameters, including sampling frequency, modulation frequency, chirp duration, chirp bandwidth, etc, were investigated in detail through experiments on a glass fiber reinforced polymer specimen with multi-scale diameter-to-depth ratio defects. This paper aims to help yield a better understanding of the physical mechanism behind TWR and propose a workable scheme for testing parameter selection in practical applications.