We propose a simulation scheme for stochastic differential equations with a reflecting boundary condition. The rate of convergence is the same as in the usual Euler-Maruyama scheme, and since our domain is a half-space or an orthant, the simulation is very easy to achieve.