Scenario based decision-making problems is that the problem complexity quickly increases as the number of scenarios increases. Scenario reduction aims at selecting a small number of scenarios to represent a large set of scenarios for decision making, so as to necessarily reduce the computational complexity while preserving the solution quality of using a large number of scenarios. A linear programming based scenario reduction method is proposed in this work. Transportation distance between the original scenario set and the selected scenario subset is used as a criterion to update the scenario selection. Reduction algorithm is proposed based on transportation distance minimization. The proposed algorithm relies on solving linear programming problems. The scenario subset updating step and the probability. Value assignment step are performed in an iterative manner until the transportation distance converges. The proposed method is applied to address scenario-based stochastic optimization problem. The application on chance constrained optimization problem shows that the proposed LP-based scenario reduction method can lead to a solution that satisfies the desired reliability level with reduced number of scenarios, while significantly reduce the computational complexity of scenario-based stochastic optimization problem.