A　blocking　lot-streaming　flow　shop　(BLSFS)　scheduling　problem　involves　in　splitting　a　job　into　several　sublots　and　no　capacity　buffers　with　blocking　between　adjacent　machines.　It　is　of　popularity　in　real-world　applications　but　hard　to　be　effectively　solved　in　light　of　many　constrains　and　complexities.　Thus,　the　research　on　optimization　algorithms　for　the　BLSFS　scheduling　problem　is　relatively　scarce.　In　view　of　this,　we　proposed　a　hybrid　discrete　artificial　bee　colony　(HDABC)　algorithm　to　tackle　the　BLSFS　scheduling　problem　with　two　commonly　used　and　conflicting　criteria,　i.e.,　makespan　and　earliness　time.　We　first　presented　three　initialization　strategies　to　enhance　the　quality　of　the　initial　population,　and　then　developed　two　novel　crossover　operators　by　taking　full　of　valuable　information　of　non-dominated　solutions　to　enhance　the　capabilities　of　HDABC　in　exploration.　We　applied　the　proposed　algorithm　to　16　instances　and　compared　with　three　previous　algorithms.　The　experimental　results　show　that　the　proposed　algorithm　clearly　outperforms　these　comparative　algorithms.　©　2018,　Springer　Nature　Singapore　Pte　Ltd.