Bearing-only passive sensors have the advantage of being non-detectable, but they come with target state observability limitations. A new approach, the unscented Gauss-Helmert filter that fuses out-of-sequence acoustic measurements (OOSM-A) and electro-optic (EO) measurements (OOSM-AE), has been developed recently to fuse non-delayed and delayed measurements from two heterogeneous passive sensors on a single platform to overcome these observability issues. In this paper, we extend the OOSM-AE approach to use interacting multiple models (IMM) to improve tar-get tracking accuracy when tracking maneuvering targets. The maneuvers considered are circular motion and S-turns. The resulting IMMOOSM-AE handles the delayed acoustic measurements as out-of-sequence measurements. Scenarios are simulated and tested with both IMMOOSM-AE and OOSM-AE and results are presented.