To realize a ubiquitous communication environment, network providers are currently attaching great importance to wireless backhaul options for the dense deployment of network nodes in urban areas. However, the increase in hop count due to wireless backhaul results in higher multi-hop delays. Network providers need an efficient UE association strategy to minimize multi-hop delays, considering the demand for future low-latency network services. To this end, we consider realistic TDMA frame configuration for downward integrated access and backhaul (IAB) communication and formulate the problem of end-to-end (E2E) delay minimization for UE association. To reduce the complexity of the problem, we apply a piecewise linear function to approximate the queueing delay part and consider an approximate optimization-based scheme ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AOPT</i> ), which can be computed offline, as a reference scheme. To make the problem solvable online, we simplify the problem and propose a Delay-based user Association Control Scheme, termed <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DACS</i> , that adaptively enables peak delay minimization by avoiding congested links and selecting routing paths with fewer hops for user association. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DACS</i> shows comparable performance to <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AOPT</i> and outperforms other comparison schemes in terms of the 98th percentile E2E tail delay, helping each user experience acceptable low-latency services.