Covariate adjustment has been a popular approach to improve precision and power in randomized trials. However, in certain trials such as vaccine trials or evaluations of new educational programs, study units often interact and influence each other’s responses, a phenomenon known as general interference. Despite recent progress in identifying and estimating treatment effects in randomized trials under general interference, there is limited work on how to properly adjust for covariates in this setting, particularly in handling potentially complex and unknown dependencies between study units. In this work, we introduce a flexible class of covariate-adjusted estimators for treatment effects under general interference. Assuming certain smoothness conditions on the response model, our estimators are consistent, asymptotically Normal, and can incorporate some existing machine learning methods. Notably, our findings suggest that ANCOVA, a common method of covariate adjustment in randomized trials, provides a consistent, asymptotically Normal, covariate-adjusted estimator of the average treatment effect under general interference. We validate our results through a simulation study using popular network models and an empirical study. The work concludes with practical guidelines for covariate adjustment in randomized trials under general interference. This is joint work with Ralph Trane (UW-Madison).