In a previous paper, we have shown that the classical definition of E+A galaxies excludes a significant number of post-starburst galaxies. We suggested that analysing broad-band spectral energy distributions (SEDs) is a more comprehensive method to select and distinguish post-starburst galaxies than the classical definition of measuring equivalent widths of (Hδ) and [O ii] lines. In this paper, we will carefully investigate this new method and evaluate it by comparing our model grid of post-starburst galaxies to observed E+A galaxies from the MORPHS catalogue. In the first part, we investigate the UV-optical-NIR (near-infrared) SEDs of a large variety in terms of progenitor galaxies, burst strengths and time-scales of post-starburst models and compare them to undisturbed spiral, S0 and E galaxies as well as to galaxies in their starburst phase. In the second part, we compare our post-starburst models with the observed E+A galaxies in terms of Lick indices, luminosities and colours. We then use the new method of comparing the model SEDs with SEDs of the observed E+A galaxies. We find that the post-starburst models can be distinguished from undisturbed spiral, S0 and E galaxies and galaxies in their starburst phase on the basis of their SEDs. It is even possible to distinguish most of the different post-starbursts by their SEDs. From the comparison with observations, we find that all observed E+A galaxies from the MORPHS catalogue can be matched by our models. However, only models with short decline time-scales for the star formation rate are possible scenarios for the observed E+A galaxies in agreement with our results from the first paper.