This poster discusses about Bayesian optimisation of prelithiation process step with mixture of discrete and continuous variables. Bayesian optimisation is an emerging technique to optimise black-box functions with reduced cost than random search. This technique is already reported as promising in many fields, including the battery research, by many groups. However, most of its applications focus on simple problems: ones with only two or three continuous input variables. Most practical problems involve multiple discrete and continuous variables to be optimised (e.g. existence of a material as a discrete variable and process time as a continuous variable). Buildng separate surrogate models per discrete variable is a simple solution, but this will lead to the increased number of required experiments. Therefore, building a surrogate model which works with mixture of discrete and continuous variables is desired. This poster will discuss challenges in such a case focusing on prelithiation. Prelithiation is a process step to lithiate anode materials before cell accembly. This will compensate for lithium consumed during formation cycles, which will result in larger capacity and longer lifetime. Our group has investigated electrochemical prelithiation of silicon as an anode active material. In this process step, some of process variables are continuous, but others are discrete because of technical limitation or their nature. We have applied some Bayesian optimisation methods from literature (e.g. B. Ru, A. S. Alvi, V. Nguyen, M. A. Osborne, S. J. Roberts, 37th ICML, PMLR 119, 2020), and will discuss our progress from the previous poster.