Laser-induced thermal waves are widely used for materials research and other applications. For low absorption optical materials the thermal waves are typically detected by collinear photothermal deflection spectroscopy (CPDS). The available theory of CPDS is based on the approximation of geometric optics, which, while being straightforward and having been extensively used, has obvious shortcomings. For example, to calculate the signal of CPDS, the probe beam size is assumed to be infinitely small. In real applications, however, the probe beam always has a finite dimension, which results in a diffraction effect and leads to unexpected experimental error. In order to overcome such shortcomings, there is a need to develop a more sophisticated theoretical model that can encompass both photothermal deflection and diffraction effects, and give precise predictions of the nature of both effects. In this paper we present a comprehensive diffraction theory suitable for treating the effect of pulsed CPDS. The work is an extension of the theoretical model previously developed for the mirage effect. It provides a detailed theoretical analysis of the laser-induced optical diffraction effect and helps to optimize the experimental parameters. It is a useful tool not only for quantitatively describing CPDS but also for the traditional thermal lens spectroscopy. With the application of the Fresnel diffraction theory, both techniques can be treated in a unified model. A summary of the theoretical modeling and experimental studies will be presented, with an emphasis on the advantages of using a near-field detection scheme for achieving the best sensitivity and stability.