Abstract: Zero Memory NonLinearity (ZMNL) and Spherically Invariant Random Process (SIRP) are two commonly used methods in K-distribution clutter simulations. An improved simulation method, which adds abranch of Gamma-distributed variable and extends the shape parameter to arbitrary positive real, is proposed to address the clutter simulation error in the conventional ZMNL method. To reduce the computation required for the conventional SIRP method, an improved method of modulation variable generation is also proposed, what avoids large computations for solving nonlinear equations and improves the simulation speed. The simulation results verify the effectiveness of the improved methods.