A multiprobe, multiple baseline design was used to evaluate the effectiveness of persuasive writing strategy instruction. Six middle school students with emotional and behavioral disabilities (EBD) received two instructional phases of Self-Regulated Strategy Development (SRSD) instruction for writing persuasive essays over 33 days of intervention. A single paragraph essay was introduced first, followed by multiple paragraph essay instruction, followed by a generalization lesson. Assessments included: (a) at least 14 essays across: baseline, post instruction phase one, phase two, maintenance, and generalization; (b) Woodcock Johnson fluency subtests; (c) a self-efficacy measure; (d) sociality validly and student interviews; and (e) time on task during instruction measures. Assessments were scored and evaluated is several ways including: (a) essays by holistic quality, length, and number of persuasive essay elements; (b) Woodcock Johnson fluency subtest at pre- and post-testing; (c) self-efficacy at pre-, post-, and maintenance testing. Findings revealed positive effects for: (a) all essay measures at post-instructional phases, maintenance, and generalization testing periods; (b), the Woodcock Johnson fluency subtest; and (c) on the self-efficacy measure. These findings replicated and extended previously conducted written expression research with middle school students with EBD. Most importantly, results revealed that instructional order of single or multiple paragraphs appeared to work equally well when findings are compared with previous research and that students with severe EBD require extensive, intensive instruction. Implications for education of students with EBD and future research are also presented.