Sparsity is one of the intrinsic properties of real-world data, thus the sparse learning models provide great opportunities to analyze the big, complex, and diverse datasets. By enforcing properly designed structured sparsity, we can integrate the specific data structures into the learning models to simplify data models and discover predictive patterns for big data applications. To address the challenging problems in current big data mining, we proposed several novel large-scale structured sparse learning models for multi-dimensional data fusion, heterogeneous tasks integration, group structured data analysis, and longitudinal feature learning. We applied these new structured sparse learning models to analyze the multi-modal brain imaging and genome-wide array data in Imaging Genomics and discover the phenotypic and genotypic markers to characterize the neurodegenerative process in the progression of Alzheimerís disease and other complex brain disorders. We also utilized the structured sparse learning models to analyze electronic medical records and predict the heart failure patientsí readmission using the first 24-hour emergency room data.