Simplified technology and low costs have spurred the use of location-detection devices in moving objects. Usually, these devices will send the moving objects' location information to a spatio-temporal data stream management system, which will be then responsible for answering spatio-temporal queries related to these moving objects. A large spectrum of research have been devoted to continuous spatio-temporal query processing. However, we axgue that several outstanding challenges have been either addressed partially or not at all in the existing literature. In particular, in this paper, we focus on the optimization of multi-predicate spatio-temporal queries on moving objects. We present several major challenges related to the lack of spatio-temporal pipelined operators, and the impact of time, space, and their combination on the query plan optimality under different circumstances of query and object distributions. We show that building an adaptive query optimization framework is key in addressing these challenges and coping with the dynamic nature of the environment we are evolving in.