Needle insertion into soft tissue has gained considerable attention in recent years in medical applications due to its ever-increasing potential in minimally invasive procedures. Steerable bevel-tip needles offer higher maneuverability independent of the insertion depth and, consequently, are preferable in many needle steering applications compared to symmetric-tip needles. However, due to the nonholonomic kinematics of the bevel-tip needle inside soft tissue, its path planning poses a considerable challenge. Though the topic of single-target path planning is rather well studied and researched, the multiple-target path-planning problem remains under-researched. In this work, we study the path-planning problem for multiple targets based on Rapidly-Exploring-Random-Tree (RRT) algorithms. These algorithms are proper candidates for intra-operative planning of needle motion due to their fast computation and simple implementations. They also work well in high-dimensional configuration spaces and under nonholonomic kinematic constraints, both of which are the characteristics of steerable bevel-tip needle motion inside soft tissue. We present two novel RRT-based path-planning approaches to steerable bevel-tip needles to reach multiple targets inside soft tissue: a 2D path planner for preoperative applications and a 3D real-time path planner for intraoperative applications. In both planners, without the needle having to completely retract and reinsert toward each separate target, the amount of tissue damage compared to the conventional sequential insertion of the needle toward each target decreases significantly. Particularly, our 3D planner works well in real-world applications where tissue and anatomical structures may vary due to tissue deformation during insertion, patient's motion, or physiological changes. In addition, our 3D planner accounts for the needle's natural curvature variation during insertion due to tissue inhomogeneity. Moreover, both of the proposed planners have real clinical applications, where the limited size of the workspace as well as the needle's limited natural curvature impose significant limitations on the needle's path-planning problem inside soft tissue. Unlike the optimization-based methods with exponential time complexity, our planners work well with as many targets as required. Simulations demonstrate the efficiency of the proposed planners in terms of minimum targeting error and decreased needle insertion length vis-a-vis the sequential insertion of the needle for each target.