Thermal reaction norms depict how temperature influences biological performances, thus also known as thermal performance curves (TPCs). Arguably, the interplay of the thermal environmental conditions and the TPC can shape the strength of intraspecific competition. Such competition can then drive the long-term evolution of the TPC. We develop a Lotka-Volterra model, using adaptive dynamics (AD), to investigate how intraspecific competition among individuals of ectotherms drives the TPC adaptation under constant versus periodically fluctuating environmental temperatures. The competition coefficient from one individual to another is assumed proportional to the ratio of their thermal performances at the current environmental temperature, with the individuals’ thermal performance being adaptive and captured by a beta distribution. Results show that, under a constant temperature, the optimal temperature of the TPC evolves to align perfectly with the environmental temperature, with the TPC breadth shrinking to zero, reflecting local adaptation to complete thermal specialisation. In fluctuating thermal environments, adaptation produces broader TPCs, with their optimal temperature potentially mismatching the average environmental temperature. When the TPC’s optimal temperature matches the average temperature, large temperature fluctuations lead to broad TPCs (thermal generalisation). Our model also shows the emergence of bimodal TPCs under rapid and large temperature fluctuations, indicating adaptation to extreme temperatures and potentially a divergence of thermal strategies within the population. Our theoretical model has demonstrated that adaptation of TPCs in periodic thermal regimes promotes the evolution of thermal generalists and possible character divergence, compared to complete thermal specialisation in constant environments.