Wearable electronics can achieve high-fidelity monitoring of pulse waveforms on the body surface enabling early diagnosis of cardiovascular diseases (CVDs). Textile-based wearable devices offer advantages in terms of high permeability and comfort. However, knitted strain sensors struggle to capture small-range deformation signals due to stress dissipation during friction and slip of yarns within the textiles. They are optimized for mechanical adaptability and adhesive capability. In this work, the stitch configurations of knitted structure are adjusted to optimize the energy dissipation ratio during deformation and waveform fitting performance. These electric-mechanical results enabled the selection of the most suitable knitted structure for the clinical diagnosis. On the other hand, the sensor's adhesion is optimized with respect to electrical-force-strain coupling and energy transfer efficiency at the interface between skin and sensor. The balance between the storage modulus and loss modulus are adjusted via the crosslinking degree of the polyacrylamide (PAAm) hydrogel network. As a result, the optimized knitted sensor enables stable collection of pulse waveforms from the radial and dorsalis pedis arteries. In human patient evaluations, the knitting-based strain sensor can distinguish patients with different potential CVD risks through extracted characteristic indicators.