This paper considers the design of a low-complexity and high-performance precoder for multiple-input multiple-output (MIMO) systems. The precoder is designed by combining both nonlinear and non-iterative processing strategies. The proposed nonlinear precoding techniques employ a nonlinear constellation precoding technique based on maximum distance separable codes at the transmitter. We propose to reduce the computational complexity in iterative-based precoding algorithms by using less complex non-iterative singular value decomposition-based joint precoder and decoder pair design. The maximum likelihood detection for the linear MIMO channel is considered. The simulation results showed that the proposed nonlinear and non-iterative precoding schemes outperform the conventional linear MIMO precoder design, even when a reduced-complexity suboptimal strategy is adopted, considering the bit error rate performance.