diff --git a/Nodlink/README.md b/Nodlink/README.md
index 36fd60eea17e26595fa2e10ff113b81ccaf32370..5473a182047b48b7e32f17919cfa0554a64d1b9f 100644
--- a/Nodlink/README.md
+++ b/Nodlink/README.md
@@ -43,7 +43,7 @@ mkdir A-SysArmor/Nodlink/Sysdig/model # for Linux data
 cp benign_file.json my/path/to/model/benign.json
 cp anomaly_file.json my/path/to/model/anomaly.json
 ```
-* precess JSON data
+* process JSON data
 ```
 $ python3 process_behavior.py --d my/path/to/model --file benign.json
 $ python3 process_behavior.py --d my/path/to/model --file anomaly.json
diff --git a/Nodlink/nonsense_detector.py b/Nodlink/nonsense_detector.py
index bb03e5bb94638b12a83472955a369b49e78b0802..2ea4ac25595dd0cd97a05c6a65f2b457327c6df9 100644
--- a/Nodlink/nonsense_detector.py
+++ b/Nodlink/nonsense_detector.py
@@ -142,7 +142,7 @@ words taken from various contemporary text corpora, (3) a set of common stop
 words, and (4) a few million strings created by randomly concatenating items
 from 2-3 (but not the real identifiers, which are left as-is).  The current
 stored results were produced after experimenting with 2-grams, 3-grams,
-4-grams and 5-grams, and and different thresholds.  The best performance
+4-grams and 5-grams, and different thresholds.  The best performance
 achieved was reached with 4-grams, and that is the value stored in the
 ngram_data.pklz pickle file in this directory.  The pickle file stores the
 values computed by the function ngram_values(); each entry is a named tuple
@@ -181,7 +181,7 @@ thresholds and parameters, then re-scoring the example strings again, and
 iterating this process until the detector function created by
 `generate_nonsense_detector()` produces good results on real and random
 strings.  A better method for finding optimal parameter values is to use a
-multiobjectve optimization algorithm.  Nostril's parameter values were
+multiobjective optimization algorithm.  Nostril's parameter values were
 initially derived manually and then fine-tuned using the NSGA-II
 (Non-dominated Sorting Genetic Algorithm) routine in
 Platypus (https://github.com/Project-Platypus/Platypus).
@@ -739,7 +739,7 @@ def test_unlabeled(input, nonsense_tester, min_length=6, sense='valid',
                    trace_scores=False, save_to=None):
     '''Test against a file or list of strings.  'nonsense_tester' is a
     function that should return True if a given string is nonsense.  'sense'
-    indicates whether each input string should be considerd to be a valid
+    indicates whether each input string should be considered to be a valid
     string, or not.  If value is 'valid', meaning the input strings are to be
     considered valid strings and not junk, then nonsense_detector(...) should
     report False for each one; if the input strings are not valid, then
diff --git a/README.md b/README.md
index bad1f24e775a21e2c9747e5615d8ed75a87c408c..85f1d87dc3aa2205a2d7a89d82d7297a151de709 100644
--- a/README.md
+++ b/README.md
@@ -4,7 +4,7 @@
 A-SysArmor聚焦操作系统安全，基于AI等技术增强系统安全能力，提高安全分析效率。
 
 
-A-SysArmor目前以实现系统入侵检测能力，主要包含以下模块：
+A-SysArmor目前已实现系统入侵检测能力，主要包含以下模块：
 
 * **NODROP**: 系统数据采集模块，负责采集系统数据，如进程创建、文件读写、网络访问等。目前主要采集系统调用信息，详细介绍参考[这里](https://gitee.com/openeuler/A-SysArmor/blob/master/NODROP/README.md)
 * **Nodlink**: 检测模块，负责分析系统数据，基于AI/ML检测系统异常行为，并生成告警图。详细介绍参考[这里](https://gitee.com/openeuler/A-SysArmor/blob/master/Nodlink/README.md)