Accurate estimates of abundance are a cornerstone for resource managers to make effective decisions for fish conservation. However, multiple sampling methods often are required to sample fish communities, and ignoring the detection process can create substantial bias in latent state parameter estimation (e.g., abundance, survival). We developed a joint-species N-mixture model that integrated snorkel, seining, and electrofishing surveys to estimate factors affecting native and non-native fish distributions in the Santa Ana River, California. We found through data integration that native Santa Ana sucker (Catostomus santaanae) and arroyo chub (Gila orcuttii) were most abundant in wide stream channels, and the abundance of both native fishes were negatively correlated with non-native largemouth bass (Micropterus salmoides). Our results highlight the power of integrating multiple data sets into a single analysis and incorporating among-species correlation into abundance modeling. Our results also highlight a pattern of native fish declines that coincides with an expanding largemouth bass population, a concern for the management of native aquatic communities within the Santa Ana River.