<p>DNA sequence design is a crucial part in DNA computing. In this field, multi-objective evolutionary algorithms (MOEA) are widely applied for generating DNA sequences and have garnered significant optimization on the quality of DNA sequences. The task is aiming to obtain a DNA sequence set, which objective values of each sequence is calculated with others. However, existing methods use the whole population to calculate objectives of one sequence, which the state of environment in evolution has big diverge with final result. Moreover, conventional framework focus on one sequence under optimizing, lacks global perspective on the whole set, which fails to fully and effectively mine the relations between objectives. To tackle with these limitations, we proposed a novel DNA sequence design framework based on MOEA, named KPIPO, which Improved Parrot Optimization with Knee Point. KPIPO provide stable evolutionary environment which ensures less distortion on optimizing direction. In addition, KPIPO focus on the whole set when optimizing one sequence. And it is flexible on choosing whom to optimize, makes the framework has clearer progress. KPIPO outperforms classical and state-of-art algorithms on objectives concluded by previous researchers. DNA sequence design has 5 objectives to optimize, which are Continuity, Hairpin, H-Measure, Similarity and Variance of Melting Temperature, exact values of our work are {0, 0, 52, 47.71, 2.38}. Compare to recent works, it obtained the same value on Continuity and Hairpin, H-Meausre and Similarity are better than others and variance of melting temperature is at the middle level.</p>