We consider significantly enhancing the angular resolution of a single-chip multiple-input multiple-output (MIMO) radar sensor equipped with only 3 transmit and 4 receive antennas for radar imaging. We adopt the slow-time code-division multiplexing (ST-CDM) transmission scheme, which is a low-cost viable option to attain a large virtual array without reducing the unambiguous Doppler range. Moreover, to significantly improve the angular resolution without increasing antenna numbers, we introduce a practical 3 × 4 sparse linear array (SLA) design. We extend the iterative adaptive approach (IAA) to the slow-time IAA (ST-IAA) algorithm, aiming to mitigate the impact of waveform separation residuals and the high angular sidelobes on ST-CDM MIMO radar imaging equipped with SLAs. Furthermore, we construct a single-chip ST-CDM MIMO radar prototype with the 3 × 4 SLA design for experimental validations. Numerical and experimental results are provided to demonstrate the enhanced imaging performance of the radar prototype when used with the ST-IAA algorithm.