Sharing　unused　private　parking　spaces　with　public　travelers　can　improve　parking　resource　utilization.　However,　a　challenge　lies　in　the　approach　to　maximize　the　utilization　of　such　parking　resources　without　time　colliding　with　the　needs　of　these　resources　between　public　travelers　and　owners　over　time.　To　address　this　challenge,　this　study　presents　a　methodology　aiming　at　efficiently　obtaining　the　optimum　matches　between　parking　supply　and　demand　with　time-window　constraints.　First,　a　framework　for　shared　parking　is　proposed　to　guide　the　development　of　the　computation　algorithms.　Based　on　this　framework,　a　mixed-integer　nonlinear　programming　model　and　its　solving　algorithms　are　then　developed　to　maximize　the　spatial-temporal　utilization.　Further,　the　feasibility　and　validity　of　the　proposed　model　and　algorithms　are　tested　by　empirical　data　collected　in　our　real-world.　The　results　indicate　that　the　non-increased　demand　durations　and　non-descending　supply　durations　(NI-ND)　heuristic　performs　the　best　under　both　small-　and　large-scale　supply　and　demand　settings　with　extremely　high　solution　efficiency.　Moreover,　the　solutions　based　on　the　NI-ND　algorithm　are　very　close　to　the　optimal　objective,　implying　that　the　proposed　shared　parking　allocation　method　based　on　the　dual　bin-packing　principle　is　an　effective　way　to　address　the　realistic　supply–demand　relationship　of　parking　resources　under　time-window　constraints.　Finally,　the　potential　of　shared　parking　is　estimated　based　on　the　statistics　of　the　number　of　private　parking　spaces　and　parking　behavior　data　in　Beijing.　©　2021　The　Author(s)