We establish a large deviation principle (LDP) for probability graphons, which are symmetric functions from the unit square into the space of probability measures. This notion extends classical graphons and provides a flexible framework for studying the limit behavior of large dense weighted graphs. In particular, our result generalizes the seminal work of Chatterjee and Varadhan (2011), who derived an LDP for Erdős-Rényi random graphs via graphon theory. We move beyond their binary (Bernoulli) setting to encompass arbitrary edge-weight distributions. Specifically, we analyze the distribution on probability graphons induced by random weighted graphs in which edges are sampled independently from a common reference probability measure supported on a compact Polish space. We prove that this distribution satisfies an LDP with a good rate function, expressed as an extension of the Kullback-Leibler divergence between probability graphons and the reference measure. This theorem can also be viewed as a Sanov-type result in the graphon setting. Our work provides a rigorous foundation for analyzing rare events in weighted networks and supports statistical inference in structured random graph models under distributional edge uncertainty.