This paper aims to investigate movement prediction of lifting task on sagittal plane using EMG signals. We used surface EMG sensor on BILH and TRLH to measure the force and position sensor (Attitude Heading Reference System; AHRS) on wrist to measure the elbow joint angle corresponding to external coordinate of musculoskeletal system. The experimental task was lifting an object on the table which varied in weight and speed. The task was mainly analyzed in time-domain and divided in three phases; pre-lifting, lifting, and holding. First, we normalized EMG signals using holding-phase EMG actuation levels instead of conventional MVC. In sequence, weight and speed classification was applied on responses of prelifting and lifting phase. This was grounded on expected characteristics such as large initiating force to lift off. Speed was classified by increasing speed of TCL (total contraction level) of pre-lifting phase (p<0.05) and weight was classified by peak TCL of lifting phase (p<0.05). Lastly, we tried trajectory estimation for the next step. Trajectory estimations for each speed and weight conditions followed trend of trajectory change, even though we used a simple linear regression method. The correlations between the estimated and measured trajectories were about 82 percent in average.