This study investigated the relative effects of four forms of task preparation on L2 speech production. One hundred forty-four Japanese speakers of English completed an oral opinion task after 10 minutes of preparation. The same task repetition condition involved four iterations of the same pedagogic task (same procedures, same content), the parallel task repetition condition involved four iterations of the same type of task (same procedures, different content), the first language (L1) planning condition involved dyadic planning for the task in Japanese, and the second language (L2) planning condition involved the same in English. The effects of these task implementation options were compared in terms of measures of conceptualization, formulation, and monitoring of L2 speech production (Kormos, 2006; Levelt, 1989) as well as overall L2 speech fluency during task performance. Results suggest that different preparatory options might support L2 learners' speech production in complementary ways by alternately reducing the conceptualization, formulation, and monitoring demands they face on tasks. Results provide an initial basis for implementing tasks to effectively support L2 learners' emerging abilities to parallel process content and language during real-time speech production (Skehan, 2009). Options for sequencing preparatory support in task-based language teaching are discussed.