The human brain undergoes rapid growth in both structure and function from infancy to adulthood through childhood and adolescence, and this significantly influences cognitive and behavioral development in later life. Developmental connectomics provides unprecedented opportunities for exploring the typical and atypical development of brain networks through non-invasive mapping of structural and functional connectivity patterns. In this talk, I will describe the computational framework of developmental connectomics and our recent works in connectome development in normal population and developmental disorders (e.g., autism). Specifically, I will highlight five fundamental principles of connectome development, including towards an optimal global balance between network segregation and integration, a development order from primary to higher-order functioning regions, different maturation modes of structural and functional connectomes, towards more stabilized and regionally differentiated modular dynamics, and the higher brain network dynamics presents high vulnerability in autism.