Functional gene networks link genes based on their functional relatedness, which is inferred from various complementary biological datasets. Gene networks comprising vast amounts of data can be used to predict which genes are associated with complex traits. Decades of studies in plant biology using the model organism Arabidopsis thaliana have generated large amounts of information, enabling the development of a system-level molecular network. AraNet (currently version 2) is a genome-scale functional gene network for Arabidopsis thaliana, constructed by integrating 19 types of genomics data and can be explored through a web-server (http://www.inetbio.org/aranet) to identify candidate genes for traits of interest. AraNet provides two alternative search paths for users to identify candidate genes and functions. The web server also exploits ortholog relationships between plant species and projects the genes of 28 other plant species (as of April, 2016) into the network of Arabidopsis genes. This allows researchers to use AraNet to predict genes/functions of not only Arabidopsis but also other non-model plants by expanding the functional knowledge of Arabidopsis. Here, we present a detailed description of how to search the AraNet network and interpret the search results to study plant gene functions and their associations with complex phenotypes.