Low temperature combustion (LTC) in diesel engines can be enabled using a multitude of fuel injection strategies, coupled with the elevated use of exhaust gas recirculation and intake boost. The common modes of LTC include the single-injection LTC with heavy EGR and the homogeneous charge compression ignition (HCCI), implemented with multiple early-injections during the compression stroke. Previous research indicates that the single-injection LTC is more suitable at low engine loads while the HCCI combustion can be targeted towards mid-load operation. To extend the load range of the LTC cycles, there is an urgent need to enable switching on-the-fly between the two combustion modes. The mode-switching is complicated by the fact that the challenges of enabling and ensuring stable engine operation under these two LTC modes are notably different. Moreover, the LTC cycles are inherently more sensitive to small changes in the operating variables such as the combustion phasing and the intake dilution, and therefore, the combustion control system must be able to adequately respond to such disturbances on a cycle-by-cycle basis.In this work, cylinder pressure measurement-based computation of combustion phasing and indicated mean effective pressure (IMEP), demonstrated in the authors' previous work for single-injection diesel LTC, has been used to enable and stabilize the LTC modes by precise control of single/multi-injection events. The IMEP estimation technique has been extended to modulate multiple fuel-injection events for dynamic load and stability control of HCCI combustion. A mode-switching algorithm is then proposed and demonstrated with engine tests, for enabling seamless transition between the two modes of LTC, by executing a pre-defined sequence triggered by an IMEP threshold, while pressure feedback-based control over individual injection events ensures the stability of the combustion. Representative results with controller gain modification indicate the possibility of controller tuning for improving the mode-switching time and transient performance. Modifications in the proposed algorithm are suggested to optimize the performance and enhance the robustness of the mode-switching process.