This paper presents a modified exponential time differencing (ETD) method for solving stiff nonlinear partial differential equations (PDEs). The original ETD method, proposed by Cox and Matthews, faced numerical instability issues, particularly when the matrix L was not diagonal. The authors propose a modification that addresses these problems using complex analysis and contour integrals in the complex plane, enabling the method to handle non-diagonal operators effectively. The modified ETD scheme is compared with other methods such as implicit-explicit (IMEX), split-step (SS), integrating factor (IF), and sliders (SL) for solving the KdV, Kuramoto-Sivashinsky, Burgers, and Allen-Cahn equations. The results show that the modified ETD scheme outperforms the others in accuracy and stability, especially for fixed time steps. The method is implemented using MATLAB for two of the equations, demonstrating its effectiveness. The paper also discusses the challenges of applying these methods to non-diagonal problems and highlights the advantages of the modified ETD scheme in terms of accuracy, stability, and computational efficiency. The study concludes that the modified ETD scheme is the best method for solving stiff PDEs in one space dimension.This paper presents a modified exponential time differencing (ETD) method for solving stiff nonlinear partial differential equations (PDEs). The original ETD method, proposed by Cox and Matthews, faced numerical instability issues, particularly when the matrix L was not diagonal. The authors propose a modification that addresses these problems using complex analysis and contour integrals in the complex plane, enabling the method to handle non-diagonal operators effectively. The modified ETD scheme is compared with other methods such as implicit-explicit (IMEX), split-step (SS), integrating factor (IF), and sliders (SL) for solving the KdV, Kuramoto-Sivashinsky, Burgers, and Allen-Cahn equations. The results show that the modified ETD scheme outperforms the others in accuracy and stability, especially for fixed time steps. The method is implemented using MATLAB for two of the equations, demonstrating its effectiveness. The paper also discusses the challenges of applying these methods to non-diagonal problems and highlights the advantages of the modified ETD scheme in terms of accuracy, stability, and computational efficiency. The study concludes that the modified ETD scheme is the best method for solving stiff PDEs in one space dimension.