Dysarthria speakers suffer from poor communication, and voice conversion (VC) technology is a potential approach for improving their speech quality. This study presents a joint feature learning approach to improve a sub-band deep neural network-based VC system, termed J-SBDNN. In this study, a listening test of speech intelligibility is used to confirm the benefits of the proposed J-SBDNN VC system, with several well-known VC approaches being used for comparison. The results showed that the J-SBDNN VC system provided a higher speech intelligibility performance than other VC approaches in most test conditions. It implies that the J-SBDNN VC system could potentially be used as one of the electronic assistive technologies to improve the speech quality for a dysarthric speaker.