Background: Patients with anaplastic lymphoma kinase-rearranged (ALK+) NSCLC inevitably acquire resistance to ALK inhibitors. We hypothesized that longitudinal monitoring of cell-free plasma DNA (cfDNA) next generation sequencing (NGS) could predict the response and resistance of TKI therapy in ALK+NSCLC Methods: Patients with ALK+ advanced NSCLC determined by standard tissue testing and planned for TKI therapy were prospectively recruited. Plasma was collected before therapy (n = 92), two months post-therapy (n = 58), and at progression (n = 35). Plasma DNA NGS analysis was done retrospectively by Guardant360. Results: From April 2015 to July 2019, 92 patients enrolled; 81 (88.0%) received ALK TKI as first-line (crizotinib, n = 59; alectinib, n = 22), 10 (10.9%) received TKI as second-line (alectinib, n = 6; crizotinib, n = 2; ceritinib, n = 1; brigatinib, n = 1), and 1 (1.1%) was treated in third-line (lorlatinib). At the cut-off date of January 28, 2020, 56 of 92 patients had disease progression. Circulating tumor DNA (ctDNA) was detected in 69 baseline samples (75%); among these were 43 ALK fusions (62.3%) and 1 ALK G1202R without fusion (1.4%). Fusions included EML4-ALK v1 (n = 19), EML4-ALK v3 (n = 14), CLTC-ALK (n = 1), TPM3-ALK (n = 1), GCC2-ALK/CLIP4-ALK (n = 1), and other EML4-ALK fusions (n = 7). Eight patients developed ALK resistance mutations after crizotinib therapy: L1196M (n = 5), G1269A (n = 1), G1202R (n = 1), and co-occurring F1174L, G1202R, and G1269A (n = 1). Two patients developed ALK resistance mutations after ceritinib: G1202R (n = 1), and co-occurring G1202R and T1151R (n = 1). The absence of detectable ctDNA at baseline was associated with longer progression-free survival (PFS; median 36.1 vs 11.6 months, HR 0.432, p = 0.004) and overall survival (OS; median not reached vs 27.9 months, HR 0.418, p = 0.034). Patients with clearance of ctDNA at two months (n = 29) had significantly longer PFS (median 25.4 vs 13.9 months, HR 0.343, p = 0.030) and OS (median not reached vs 25.7 months, HR 0.173, p = 0.035) than those without clearance (n = 22). Patients with co-occurring TP53 alterations and ALK fusions at baseline (n = 9) showed shorter PFS (median 7.0 vs 12.5 months, HR 3.596, p = 0.0154) than those without TP53 alterations (n = 35). Conclusions: NGS of cell-free plasma DNA is useful not only for the detection of ALK fusions and resistance mutations but also for assessing prognosis and monitoring the dynamic changes of genomic alterations in ALK+ NSCLC treated with ALK TKI.