Wind data is critical for aircraft safe operation and air traffic management. Especially during the transition and reverse transitions of vertical takeoff and landing UAVs dynamic information of wind-field is vital. This paper proposes a 1-D convolutional neural network-based dynamic estimation of the wind-field for a UAV. Simulation is carried out for the generation of trim input data for different wind and trim conditions which is based on analytical approach. Nonlinear constrained optimization is performed to determine trim input. Further, the set of control inputs and trim states are used to train and validate the proposed neural network.