This paper presents a joint decoding scheme for block coded signals in an overloaded multiple-input multiple-output (MIMO)-OFDM system. In a normal MIMO-OFDM system, the number of the transmit antenna is the same as that of the received antenna. In the conventional MIMO system, symbol detection is carried out over the signals received by multiple antenna elements and is separated from decoding. However, in order to increase the throughput of the system, a more number of antenna elements is required, which leads to larger power consumption owing to additional radio frequency (RF) units. Instead of splitting detection and decoding, in this paper, a joint maximum likelihood (ML) decoding scheme in an overloaded MIMO-OFDM system is presented. Based on ideal interleaving, independence among coded symbols spread over subcarriers is assumed. As an example of the proposed system, bit error rate performance of Hamming coded and spatially multiplexed signals with the joint ML decoding on a Rayleigh fading channel is investigated. Furthermore, a complexity reduction scheme of the joint ML decoding is also evaluated through computer simulation and bit error performance with the proposed complexity reduction scheme is presented.