The thermal limit of overhead lines is one of the key constraints in system scheduling problem to ensure the system security and efficiency. In order to achieve better utilization of overhead lines, especially in a distribution system with high renewable energy penetration, this paper proposes a stochastic system scheduling model with Dynamic Line Rating (DLR) constraints. The nonlinear nonconvex DLR constraints derived from the Heat Balance Equation (HBE) are relaxed to Second-Order Cone (SOC) formulation. Therefore, mixed-integer SOC Programming (SOCP) is developed, which is more accurate than the traditional linear programming and can be solved efficiently. A modified IEEE 14-bus system is applied to investigate the effectiveness of relaxation and demonstrate the economic benefit.