The purpose of this study was to develop the analysis algorithm that is able to estimate the baseball bat trajectory from swing beginning to follow-through phase in the practice swing motion and to examine the accuracy of the proposed method using the 3D motion capture system. The measurement system in this study, which consists of inertial sensor system, measured 3-axis angular velocity and 3-axis acceleration. To compensate the error of velocity during follow-through phase, the backward direction integration from the instance of swing stop was performed. The sensor fusion using the Adaptive Kalman filter by compensating velocity compensated the error of acceleration integration during follow-through phase, and it estimated the 3D bat trajectory in a global coordinate by combining the sensor fusion and the compensation using motion characteristics. The measurement experiment was conducted to indicate the accuracy of the proposed method using 3D motion capture system. The 3D bat trajectory from swing beginning to follow-through phase estimated by the proposed method corresponded with the 3D bat trajectory obtained by the 3D motion capture system. This result indicates that the error by acceleration integration during follow-through phase is compensated. Therefore, the effectiveness of the proposed method was indicated and the proposed method can be used to evaluate baseball swing include follow-through phase with high accuracy.