An apparatus and a method of arbitrarily adjusting the output time of caption information included in a received multimedia transport stream for broadcasting, by user input. The method includes: receiving an input of a time value for adjusting the output time of caption information, from a user through a user interface; by demultiplexing the multimedia transport stream, dividing the transport stream into video, audio, and data packets, decoding each packet, and storing the decoded packets in buffers corresponding to a type of each packet; resetting the output time of the caption information by adding the input time value to or subtracting the input time value from the output time included in the video or audio packet; and according to the output time of the reset caption information, blending the data stored in the buffer with the video information and presenting the blended information on the screen.