This research aims to illustrate the effectiveness of short story writing learning with modeling methods in private high school students of The Langkat Bahorok Regional Government compared to lecture methods. The population in this study was all students of class VIII of the Private High School Langkat Bahorok With a sample number of 20 people for each treatment. A research method is an experiment that compares between two groups given different treatment, in which case the experimental group is given short story writing learning with modeling methods and the control group is given learning by lecture method. From the distribution results obtained: (1) The experimental group that got learning to write short stories with modeling methods had an average value (Mean) of 74.75 and standard deviation (SD) of 15.66; (2) The control group that learned to write short stories with lecture methods had an average score (mean) of 62.5 and Standard Deviation (SD) of 10.97. Furthermore, from the calculation of the hypothesis test obtained greatly t hitung = 2,37 at a significance level of 5% price t tabel = 2,02 and to a degree 1% price t tabel = 2,71. Because of the price t hitung Bigger than t tabel (2,02 < 2,37 < 2,71), then the nil hypothesis (Ho) can be rejected and the alternative hypothesis (Ha) is accepted, which means there is a significant influence between the learning outcomes of writing short stories with modeling methods and learning outcomes of writing short stories with lecture methods on students Private High School Langkat Bahorok.