In this thesis, we proposed a goal-oriented navigation reinforcement learning network called GRU- Attention based TD3 network, which takes Lider measurements, the distance between target position, and yaw toward the target as state inputs. The policy in the network will output continuous action: forward velocity and yaw angular velocity. Our proposed network can perform obstacle avoidance navigation without prior knowledge of the environment. We train our network in a simulation environment. To show that our proposed network is better in navigation tasks, we compare the performance with two other networks: the pure TD3 network and the GRU-based TD3 network in multiple simulation worlds. The experiments show that our proposed network can bypass the obstacles safely and arrive at the goal positions as fast as possible.