The use of minimally invasive procedures for breast tumor diagnosis and treatment, such as needle biopsy and radiofrequency ablation (RFA), is steadily increasing. Accurate needle insertion requires solving the problems of tissue deformation and target displacement. In this study, we developed a robotic needle insertion method to improve the precision of diagnostic biopsy and RFA treatment. The mechanical probe was designed to reduce tissue displacement by pressing the breast tissue before needle insertion: a technique that is known as preloading. We focused on the needle insertion phase and evaluated the insertion accuracy achieved. Using a numerical simulation model and an actual hog breast, we compared tissue preloaded needle insertion with normal needle insertion. The data obtained with both test systems showed that targeting errors were greatly reduced using preloading-based needle insertion, as compared with normal needle insertion. The procedure is expected to offer a safe and effective alternative to the traditional methods of needle insertion for breast tissue biopsy or RFA. Our study also revealed the relationship between insertion accuracy and preloading probe force.