Multivariate statistics are commonly used by ecologists to study spatial and temporal community dynamics to better inform management decisions. Since these methods are a universal tool to analyze data, it is important to assess their effectiveness using long-term datasets in well-studied systems. The objectives of this study were to identify trends in the fish communities and to characterize the community trajectory (i.e., directional or nondirectional) of the Illinois River, a large tributary to the Mississippi River, using nonmetric multidimensional scaling (NMDS). Although NMDS has been used to evaluate community trajectory in multivariate space over time in smaller rivers, there has been less attention given to larger rivers, which may exhibit different dynamics. To evaluate changes in large river fish communities, we used a long-term electrofishing monitoring dataset (1957–2013) that includes fish abundance data from six reaches of the Illinois River, Illinois, U.S.A. Temporal changes in Illinois River fish communities were evident in NMDS ordinations. Larger changes in NMDS ordinations were associated with improving water quality conditions, whereas the invasion of bigheaded carp (silver carp Hypophthalmichtys molitrix and bighead carp H. nobilis) produced smaller and less variable changes. Across all six reaches of the Illinois River, the trajectory of fish communities was directional in multivariate space, with no return to an initial or predisturbance condition. Community trajectory in larger and more heavily-modified rivers may be substantially different and more complex than that of smaller and less heavily modified rivers.