In the absence of sufficient vaccination and rapid mutation of the virus, joint interventions of detection, contact tracing and social distancing are considered to be key strategies in suppressing emergent epidemics. Here, we propose the susceptible-exposed-infected-quarantined-recovered (SEIQR) model with detection, contact tracing, and dynamical social distancing to quantify the effectiveness of joint interventions. Under the framework of time-varying networks, we analytically derive the epidemic threshold and the effective reproductive number under joint interventions by using the mean-field approach. Experimental results show that detection and contact tracing can significantly reduce the epidemic scale, while social distancing can remarkably delay the peak time and suppress the second outbreak of the epidemic. Compared with constant social distancing, the effectiveness of dynamical social distancing depends on the starting time and the intervention period, and extensive detection can greatly reduce its dependence on them. Further, we explore the effects of heterogeneous distributions of physical distance and contact duration on the effectiveness of joint interventions in high-resolution empirical contact networks. We find that extensive contact tracing is more effective in decreasing the “false negatives,” but at the expense of unnecessary large-scale quarantines of the “false positives.” Meanwhile, the early implementation of long-term social distancing can effectively reduce the “false negatives” and “false positives.” In conclusion, it is very important to ensure extensive detection and contact tracing, while maintaining moderate social distancing to suppress the epidemic spread and prevent epidemic recurrence.