This study aims to determine whether the application of the snowball throwing learning method can improve the cognitive learning outcomes of class X students on the material of Lithospheric Dynamics at SMA Negeri 61 Jakarta. This research is included in quasy experimental research with a non-equivalent control group design that uses a control class and an experimental class. The sampling technique in this study used purposive sampling method. The samples used in this study are class X-5 as an experimental class that will use the snowball throwing learning model in the learning process and class X-6 as a control class that will use conventional learning methods (lectures) in the learning process. The data obtained comes from the pre-test and post-test question scores which refer to the revised Bloom's taxonomy cognitive ability levels. The techniques used to analyze the data in this study are instrument test, data test as analysis requirements, and data analysis test. Based on the test results, the instrument was declared valid and reliable, and the data was declared normal and homogeneous. Based on data testing taken in the field, the results showed that there was an increase in students' cognitive learning outcomes in the experimental class. The experimental class experienced a significant increase in cognitive learning outcomes with the acquisition of an average value to 90.06 from 66.03. While the acquisition of the control class from 62.47 to 80.67. This conclusion is also supported by the application of the snowball throwing learning method can improve students' cognitive learning outcomes in class X on the material of Lithospheric Dynamics at SMAN 61 Jakarta.