Optical lenses are extensively used to enhance the performance of light-emitting diodes. Both uniformity and efficiency are important performance indicators in lens design; however, improving uniformity always lowers efficiency. In this study, the Taguchi method and principal component analysis (PCA) are integrated to optimise the lens shape for two quality objectives, namely, uniformity and efficiency. The Taguchi method was conducted twice to establish the signal/noise ratio of the two quality characteristics for calculating the principal components in PCA. Then, the optimum parameters obtained by the Taguchi method were processed by PCA. The correlated individual responses were converted to the principal components which explained most of the dataset and were considered as the single quality characteristic for the optimisation. The combined method resolved the difficulties of optimising multiple quality characteristics without sacrificing any particular quality characteristic while the traditional Taguchi method can only be applied to the single quality characteristic. A LED light source fitted with a secondary lens designed by the proposed method showed over 92% light efficiency and an improvement in uniformity.