Along with the evolution of heterogeneous multi-core architectures, advanced thread scheduling algorithms for maximizing system efficiency are needed. As power density increases with technology, core temperature variations can cause bottlenecks in the system performance. More efficient thread scheduling algorithms for heterogeneous multi-core architectures and advanced dynamic thermal management (DTM) techniques are needed in the current multi-core era.The performance and power requirements of a thread vary significantly during its execution. Threads can be reassigned to cores upon detection of such variations. This can improve the power efficiency and performance of the heterogeneous multi-core architectures. Traditional online learning schemes rely on sampling to determine the thread-to-core affinity. However, as the number of cores increase, the sampling overhead would be high. Moreover, the core temperature must be below a safe level, or else significant power dissipation can occur. Overheating of the cores can cause hot spots which severely reduce their lifespan and may affect the runtime performance. The objective of this thesis is to analyze a dynamic thermal management technique, which considers core temperature variations for heterogeneous multi-core architectures. While doing so, the thread scheduling decisions in heterogeneous multi-core architectures are taken using an online estimation technique. A scheduler is proposed which can estimate the power and performance of the current thread on all the cores. Depending upon the expected power and performance requirements, the appropriate thread-to-core pair is selected by the scheduler. A thermal profiler is proposed which classifies preempted threads into hot, warm and cold categories. These threads are classified according to the temperature variations which they produce on any particular core. An additional task of rescheduling these classified threads is performed by the scheduler, considering the core temperatures. The queuing network model is used to analyze the proposed design.