Innovations in engineered wood products have led to an increased interest in the design and construction of high-rise timber buildings. However, concerns arise when considering the fire safety of tall timber structures. Similar to other types of structures, the fire resistance (FR) of these types of buildings is greatly affected by the performance of its connections along with the performance of the main structural members. Previously, most of experimental studies focused on the fire-performance (FP) of the metal connections loaded in tension, and parallel to the grain of timber elements, while they were exposed to the standard fire curves. However, contributions in a case study on the fire-performance of tall timber structures confirmed that for the contemporary tall timber buildings, connections would be loaded in different directions and the fire behavior of timber structures directly would be influenced by the percent of exposure of the combustible elements to the fire. Therefore, testing assemblies by prescriptive rules and standard fire curves could not be a precise assumption. In ambient temperature, connections loaded parallel and perpendicular to the grain perform differently. Brittle splitting failure is the dominant failure mode for the connections loaded perpendicular to the grain. Considering the different failure modes at normal temperature and the effect of metal fasteners in increasing the charring in the connection location, raised a question on how the loading direction will affect the fire-performance of doweled connections. In this study heavy timber beam-to-girder and beam-to-wall steel and aluminum doweled connections were tested at three different thermal conditions. Tests were first performed at ambient temperature where specimens were loaded perpendicular to the grain to determine the capacity of the assembly. Then post-fire-performance (PFP) tests were conducted where the assemblies were exposed to a non-standard fire, allowed to be cooled, and then loaded perpendicular to the grain to investigate the residual capacity of the connection. Finally, the assemblies were loaded perpendicular to the grain while they were exposed to a non-standard fire curve in order to study the fire-performance (FP) of the assemblies. Parameters such as width of the gap between the beam and the girder or wall, beam thickness around the connection, and the thermal material properties influence the fire resistance. A framework to model timber connections at room temperature and elevated temperature, based on a coupled finite-element (FE) heat-displacement model, was developed and applied in a parametric study of the previously tested connection assemblies. Comparison with experimental results showed that the FE models provided good estimates of measured temperatures and the load-carrying capacity at normal temperature and in fire.