The optimization of heat conduction is a critical task with widespread applications, and the approaches are typically categorized into two main categories: thermal conductivity distribution optimization (TCDO) and heat source layout optimization (HSLO). While extensive research efforts have been devoted to each of these two categories, standalone TCDO and HSLO limit the design possibilities and may lead to suboptimal solutions. In this work, a collaborative methodology combining TCDO and HSLO is proposed by transforming the collaborative optimization problem into a two-level nested problem. In this approach, TCDO forms the inner subproblem, tackled using the gradient-based method, while HSLO constitutes the outer subproblem addressed through Bayesian optimization (BO). The proposed method is employed to solve two problem cases involving volume-to-point and volume-to-edge boundaries, respectively. The results demonstrate that the present method is capable of achieving collaborative optimization of TCDO and HSLO for both scenarios of continuous and discrete thermal conductivity distributions. Comparing with standalone TCDO and HSLO that reduce the average temperature by ▪ and ▪, respectively, the proposed method achieves a significantly greater reduction of ▪, underscoring its efficacy. We anticipate that the proposed method will serve as a valuable tool for optimizing heat conduction across diverse applications, and its adaptable framework holds promise for addressing broader optimization challenges. •Thermal conductivity and heat source layout are collaboratively optimized.•A nested optimization framework is proposed for the collaborative optimization.•Collaborative optimization outperforms standalone ones by 38.26 K and 82.55 K.•The proposed method is adaptive and efficient for wider range of problems.