Inadvertent cuts to blood vessels and nerves poses a significant risk during percutaneous needle procedures, often leading to serious injuries and even death. We propose a computer-assisted photoacoustic imaging-based device that is able to detect these vascular structures and robotically guide the surgeons in avoiding them. A fiber-coupled pulsed laser diode capable of generating photoacoustic signals is attached through a ferrule, where a 2.5 mm diameter ultrasound ring transducer receives the corresponding photoacoustic waves. The integrated device is secured on an XYZ axis linear translational stage configuration, and robotically navigated through vessel-modelling phantoms to reach a targeted region of interest. A steering feedback algorithm calculates the relative position of the device with respect to each vessel, generates a 2D map of the navigational plane, and controls the stages to steer the device accordingly towards the target while avoiding the vessels. We first ran the algorithm in a water phantom to demonstrate feasibility, and then in a milk solution to model real tissue scattering. Our proposed device successfully avoids the phantom blood vessels in both cases through photoacoustic detection, and the corresponding 2D navigational path and plane through the phantom is mapped and recorded. Our results demonstrate that a computer-assisted photoacoustic imaging-based device is a viable method of intraoperatively guiding percutaneous needle procedures. The ability of our proposed guidance device to detect and avoid damage to blood vessels and nerves can further be used to optimize biopsies and tumor removal procedures in various parts of the body.